CN105049849A

CN105049849A - Image decoding device

Info

Publication number: CN105049849A
Application number: CN201510489608.XA
Authority: CN
Inventors: 山本智幸
Original assignee: Sharp Corp
Current assignee: Has Media International Ltd
Priority date: 2010-02-24
Filing date: 2010-12-27
Publication date: 2015-11-11
Anticipated expiration: 2030-12-27
Also published as: USRE48074E1; BR112012021050A2; JP2016103851A; JPWO2011104993A1; EP2541938A1; CN108810532B; US9832465B2; CN105049860A; CN105049849B; HK1217255A1; JP2017079473A; HK1217253A1; CN105141959A; MX2019005605A; CN102771126A; MX2012009474A; JP2014233094A; BR112012021050A8; HK1217254A1; CN105141959B

Abstract

The invention provides an image encoding device and an image decoding device. The image encoding device includes a first prediction parameter determination section (53) for selecting, for each of prediction units belonging to a first group, a prediction parameter from a basic set; a second prediction parameter determination section (55) for selecting, for each of prediction units belonging to a second group, a prediction parameter from a reduced set (i) including at least a part of the prediction parameter(s) selected by the first prediction parameter determination section (53) and (ii) is constituted by a prediction parameter(s), the number of which is not more than the number of prediction parameters included in the basic set; and a prediction parameter encoding section (243) for encoding (i) information indicating which one of prediction parameters is selected by the first prediction parameter determination section (53) and (ii) information indicating which one of prediction parameters is selected by the second prediction parameter determination section (55).

Description

Picture decoding apparatus

The divisional application that the application is the applying date is on December 27th, 2010, application number is 201080064667.8, denomination of invention is the application for a patent for invention of " picture coding device and picture decoding apparatus ".

Technical field

The present invention relates to encodes to image generates the picture coding device of coded data.In addition, the picture decoding apparatus that the coded data that generates is decoded to using such picture coding device is related to.

Background technology

In order to transmission efficiently or record moving image, employ dynamic image encoding device.As concrete moving image encoding mode, include, for example H.264/MPEG-4AVC the mode etc. adopted in joint development codec in (non-patent literature 1) and VCEG (Video Coding Experts group) and KTA software.

In such coded system, the image (picture) forming moving image by by the section obtained Image Segmentation Using, section to be split and hierarchy that the macro block that obtains and the sub-block splitting macro block and obtain are formed manages, normally encode by every sub-block.

In addition, in such coded system, usually, generate predicted picture based on the local decoder image obtained by carrying out coding/decoding to input picture, and the differential data between this predicted picture and input picture is encoded.In addition, as the generation method of predicted picture, be called that the method for inter prediction (interprediction) and infra-frame prediction (intraprediction) is known.

In inter prediction, the motion compensation of motion vector by the reference picture applications exploiting to frame entirety being decoded in the reference frame that obtains, generates the predicted picture in forecasting object frame.In addition, in inter prediction, predicted picture can also be generated with reference to multiple reference picture, now, use weight coefficient is multiplied by the pixel value of each reference picture and the value obtained to generate predicted picture.

On the other hand, in infra-frame prediction, based on the local decoder image in same frame, generate the predicted picture in this frame successively.Specifically, in infra-frame prediction, usually, by component unit region (such as, macro block) each prediction unit (such as, sub-block), select arbitrary prediction direction among prediction direction contained from prespecified prediction direction (predictive mode) group, and along selected prediction direction, extrapolation is carried out to the pixel value of the reference pixel in the decoded picture of local, generate the predicted pixel values on this forecasting object region thus.

So, predicted picture generally can generate based on the Prediction Parameters of motion vector, weight coefficient or predictive mode etc.

Look-ahead technique document

Non-patent literature

Non-patent literature 1:ITU-TRecommendationH.264 (11/07) (in November, 2007 is open)

The problem that invention will solve

But, in order to suitably generation forecast image in moving image decoding apparatus, need to encode to the Prediction Parameters used in dynamic image encoding device, and be sent to moving image decoding apparatus, the problem that the code amount that therefore there is coded data increases because of Prediction Parameters.

Such as, in above-mentioned existing infra-frame prediction, need for each forecasting object region, above-mentioned differential data of not only encoding, also coded representation have selected information and the prediction mode information of any predictive mode, and the code amount that therefore there is coded data increases such problem because of prediction mode information.

Summary of the invention

The present invention proposes in view of the above-mentioned problems, its object is to, realize the picture coding device cutting down the code amount being used to specify Prediction Parameters under the prerequisite of not sacrificing code efficiency and the picture decoding apparatus can decoded to the coded data generated by such picture coding device.

For solving the means of problem

In order to solve above-mentioned problem, picture coding device involved in the present invention is a kind of picture coding device of encoding to the difference of input picture and predicted picture, the feature of described picture coding device is, comprise: taxon, predicted picture is divided into multiple unit area by it, and multiple prediction units contained in constituent parts region are categorized as the 1st group or the 2nd group; 1st selected cell, it is among the baseset be made up of prespecified Prediction Parameters, selects the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 1st group; 2nd selected cell, it is among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset comprised the Prediction Parameters selected by above-mentioned 1st selected cell, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group; With Prediction Parameters coding unit, it have selected which Prediction Parameters and above-mentioned 2nd selected cell to above-mentioned 1st selected cell for each prediction unit belonging to above-mentioned 1st group and have selected which Prediction Parameters for each prediction unit belonging to above-mentioned 2nd group and encode.

According to above-mentioned such picture coding device formed, be among the reduced set formed at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset in the Prediction Parameters selected of each prediction unit of the 1st group same unit area from comprising by above-mentioned 1st selected cell for belonging to the 2nd group, select Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group, and which Prediction Parameters be have selected to above-mentioned 2nd selected cell encode.

At this, due to the Prediction Parameters for each prediction unit generally to exist relevant between the Prediction Parameters for the prediction unit near this prediction unit, be also therefore that the possibility of suitable Prediction Parameters is high for belonging to each prediction unit of above-mentioned 1st group and the Prediction Parameters selected to each prediction unit belonging to above-mentioned 2nd group.That is, for each prediction unit belonging to above-mentioned 2nd group, the Prediction Parameters selected among above-mentioned reduced set is that the possibility of suitable Prediction Parameters is high.Therefore, according to above-mentioned formation, code efficiency can not be made to reduce carry out the coding of Prediction Parameters.

In addition, in above-mentioned formation, above-mentioned reduced set comprises the reduced set at least partially in the Prediction Parameters selected by above-mentioned 1st selected cell, and the set be made up of the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset, therefore can cut down for representing for each prediction unit belonging to above-mentioned 2nd group and have selected the code amount of the information of which Prediction Parameters.

Therefore, according to above-mentioned formation, the code amount being used to specify Prediction Parameters can be cut down under the prerequisite of not sacrificing code efficiency.

In addition, picture coding device involved in the present invention is a kind of picture coding device of encoding to the difference of input picture and predicted picture, it is characterized in that, comprise: selected cell, among the reduced set at least partially of its Prediction Parameters of specifying from the generation method of the predicted picture the prediction unit comprised for completing the coding be arranged near this prediction unit, select the Prediction Parameters for specifying the generation method of the predicted picture in each prediction unit; With Prediction Parameters coding unit, which Prediction Parameters it have selected to above-mentioned selected cell for each prediction unit is encoded.

Generally speaking, for each prediction unit Prediction Parameters to exist between the Prediction Parameters for the prediction unit near this prediction unit relevant.Therefore, in the generation of the predicted picture in this prediction unit, the possibility that above-mentioned reduced set comprises Prediction Parameters is the highest.In addition, due to above-mentioned reduced set forming at least partially by the Prediction Parameters for the prediction unit be arranged near this prediction unit, the number of therefore contained in above-mentioned reduced set Prediction Parameters is less than the number of Prediction Parameters contained among the parameter set that is made up of the Prediction Parameters for the prediction unit beyond this prediction unit.

Therefore, picture coding device involved in the present invention, can generated code amount is few under the prerequisite of not sacrificing code efficiency coded data by taking above-mentioned formation.

In addition, picture decoding apparatus involved in the present invention be a kind of to by by the difference of original image and predicted picture, with represent have selected which Prediction Parameters in multiple Prediction Parameters of the generation method being used to specify predicted picture for each prediction unit selection information together with encode and decoding device that the coded data that obtains is decoded, it is characterized in that, comprise: taxon, its by form multiple unit areas of predicted picture each in contained multiple prediction units be categorized as the 1st group or the 2nd group; 1st selected cell, its reference pin is to the selection information of each prediction unit belonging to above-mentioned 1st group, among the baseset be made up of prespecified Prediction Parameters, select the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of the 1st group; With the 2nd selected cell, its reference pin is to the selection information of each prediction unit belonging to above-mentioned 2nd group, among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset comprised the Prediction Parameters selected by above-mentioned 1st selected cell, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of the 2nd group.

According to above-mentioned such picture decoding apparatus formed, can be among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset in the Prediction Parameters selected of each prediction unit of the 1st group same unit area from comprising by above-mentioned 1st selected cell for belonging to the 2nd group, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group.

At this, due to the Prediction Parameters for each prediction unit generally to exist relevant between the Prediction Parameters for the prediction unit near this prediction unit, be also therefore that the possibility of suitable Prediction Parameters is high for belonging to each prediction unit of above-mentioned 1st group and the Prediction Parameters selected to each prediction unit belonging to above-mentioned 2nd group.Therefore, according to above-mentioned formation, code efficiency can not be made to reduce the selection information less according to code amount carry out the decoding of Prediction Parameters.

In addition, picture decoding apparatus involved in the present invention is a kind of difference to passing through input picture and predicted picture, with represent have selected which Prediction Parameters in multiple Prediction Parameters of the generation method being used to specify predicted picture for each prediction unit selection information together with to encode and the coded data that obtains carries out the picture decoding apparatus of decoding, it is characterized in that, comprise: selected cell, it is with reference to above-mentioned selection information, among the reduced set at least partially of the Prediction Parameters of specifying from the generation method of the predicted picture the prediction unit comprised for completing the decoding be arranged near this prediction unit, select the Prediction Parameters for specifying the generation method of the predicted picture in each prediction unit.

Generally speaking, for each prediction unit Prediction Parameters to exist between the Prediction Parameters for the prediction unit near this prediction unit relevant.Therefore, above-mentioned reduced set comprises the possibility of the optimal Prediction Parameters of generation of the predicted picture in this prediction unit high.In addition, due to above-mentioned reduced set forming at least partially by the Prediction Parameters for the prediction unit be arranged near this prediction unit, the number of therefore contained in above-mentioned reduced set Prediction Parameters is less than the number of Prediction Parameters contained among the parameter set that is made up of the Prediction Parameters for the prediction unit beyond this prediction unit.

Therefore, the picture coding device with the formation corresponding with above-mentioned formation can generated code amount is few under the prerequisite of not sacrificing code efficiency coded data.

The picture decoding apparatus with above-mentioned formation can the coded data few to code amount like this be decoded.

In addition, the data structure of coded data involved in the present invention is a kind of difference to passing through input picture and predicted picture, with represent have selected which Prediction Parameters in multiple Prediction Parameters of the generation method being used to specify predicted picture for each prediction unit selection information together with encode and the data structure of coded data that obtains, it is characterized in that, comprise selection information, namely, in the picture decoding apparatus that above-mentioned coded data is decoded, in order to select Prediction Parameters for specifying the generation method of the predicted picture in each prediction unit among the reduced set at least partially of Prediction Parameters of specifying from the generation method of the predicted picture in the prediction unit comprised for completing the decoding be arranged near this prediction unit and the selection information of reference.

Therefore, the coded data with above-mentioned formation is the coded data reducing code amount under the prerequisite of not sacrificing code efficiency.

Invention effect

As mentioned above, picture coding device involved in the present invention is a kind of picture coding device of encoding to the difference of input picture and predicted picture, comprise: taxon, predicted picture is divided into multiple unit area by it, and multiple prediction units contained in constituent parts region are categorized as the 1st group or the 2nd group; 1st selected cell, it is among the baseset be made up of prespecified Prediction Parameters, selects the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 1st group; 2nd selected cell, it is among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset comprised the Prediction Parameters selected by above-mentioned 1st selected cell, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group; With Prediction Parameters coding unit, it have selected which Prediction Parameters and above-mentioned 2nd selected cell to above-mentioned 1st selected cell for each prediction unit belonging to above-mentioned 1st group and have selected which Prediction Parameters for each prediction unit belonging to above-mentioned 2nd group and encode.

According to above-mentioned such picture coding device formed, the code amount being used to specify Prediction Parameters can be cut down under the prerequisite of not sacrificing code efficiency.

Accompanying drawing explanation

Fig. 1 is the block diagram of the formation of moving image decoding apparatus involved by execution mode.

Fig. 2 is the block diagram of the formation representing the MB lsb decoder that the moving image decoding apparatus involved by execution mode possesses.

Fig. 3 is the block diagram of the formation representing the Prediction Parameters lsb decoder that the moving image decoding apparatus involved by execution mode possesses.

Fig. 4 is the figure of the action of the group detection unit possessed for illustration of Prediction Parameters lsb decoder.16 sub-blocks a () ~ (b) shows in macro block contained by are classified as the situation of any one in the 1st group and the 2nd group separately based on sorting technique A, c () ~ (d) shows the situation that each sub-block is classified based on sorting technique B, (e) ~ (f) shows the situation that each sub-block is classified based on sorting technique C.

Fig. 5 is by the intra prediction mode used in the infra-frame prediction in H.264/MPEG-4AVC specification, the figure represented together with the call number of giving each predictive mode.

Fig. 6 is the figure of the action of the reduced set leading-out portion possessed for illustration of Prediction Parameters lsb decoder.A () is the flow chart of the 1st example of the generation action of the reduced set represented in reduced set leading-out portion, b () is the flow chart of the 2nd example of the generation action of the reduced set represented in reduced set leading-out portion, (c) is the flow chart of the 3rd example of the generation action of the reduced set represented in reduced set leading-out portion.

Fig. 7 is the flow chart of an example of the flow process of decoding process in the 2nd Prediction Parameters lsb decoder representing that Prediction Parameters lsb decoder possesses.

Fig. 8 is the figure of other the configuration example for illustration of Prediction Parameters lsb decoder.A () is the flow chart of the generation action representing the reduced set performed by reduced set leading-out portion, (b) shows an example in neighbouring sub-block region.

Fig. 9 is the figure of the generating process of the predicted picture performed for illustration of the predicted picture generating unit that possessed by MB lsb decoder, shows the pixel as each pixel of the forecasting object sub-block of 4 × 4 pixels and the periphery of this forecasting object sub-block.

Figure 10 is the block diagram of the formation of the dynamic image encoding device represented involved by execution mode.

Figure 11 is the block diagram of the formation representing the MB coding unit that the dynamic image encoding device involved by execution mode possesses.

Figure 12 is the block diagram of the formation representing the Prediction Parameters determination section that MB coding unit possesses.

Figure 13 is the figure of the action of the Prediction Parameters determination section possessed for illustration of MB coding unit.A () shows in each sub-block of the 1st Prediction Parameters determination section for formation macro block MB the example of each sub-block belonging to the 1st group and the predictive mode selected, b () shows an example of the reduced set generated by reduced set leading-out portion when having been carried out providing as Prediction Parameters by each predictive mode shown in (a), (c) shows the 2nd Prediction Parameters determination section for belonging to each sub-block of the 2nd group and the example of predictive mode selected among the reduced set shown in (b).

Figure 14 is the block diagram of the formation representing the Prediction Parameters coding unit that MB coding unit possesses.

Figure 15 is the figure of the bit stream structure of each macro block representing the coded data referenced in the moving image decoding apparatus involved by execution mode that the dynamic image encoding device involved by execution mode generates.

Figure 16 is the figure of other the example representing basic parameter collection.A () shows an example of the parameter set paying attention to horizontal direction, (b) shows an example of the parameter set paying attention to vertical direction.

Embodiment

(moving image decoding apparatus)

The formation of the moving image decoding apparatus (picture decoding apparatus) 1 involved by execution mode is described with reference to Fig. 1 ~ Fig. 9.Moving image decoding apparatus 1 is that wherein a part employs the dynamic image encoding device of the technology adopted in H.264/MPEG-4AVC specification.

Moving image decoding apparatus 1 in brief, is generate by decoding to inputted coded data #1 and export the device of decoded picture #2.

In addition, unit area on image shown in coded data #1 is divided into multiple forecasting object region (prediction unit) by moving image decoding apparatus 1, use each and the predicted picture that generates by this forecasting object region, generate decoded picture #2.

Although below the situation of above-mentioned forecasting object region as the sub-block in macro block to be described by the macro block of above-mentioned unit area in H.264/MPEG-4AVC specification, the present invention is not limited thereto.Such as, above-mentioned unit area can be set to the region larger than macro block, can also be set to multiple macro block and repeat such region.

Fig. 1 is the block diagram of the formation representing moving image decoding apparatus 1.As shown in Figure 1, moving image decoding apparatus 1 possesses: variable length code demultiplexing portion 11, header information lsb decoder 12, MB configuration part 13, MB lsb decoder 14 and frame memory 15.

The coded data #1 inputing to moving image decoding apparatus 1 is inputted to variable length demultiplexing portion 11.Variable length demultiplexing portion 11 is by carrying out demultiplexing to inputted coded data #1, coded data #1 is separated into the coded data relevant to header information and header coded data #11a and the coded data relevant with macro block (unit area) and MB coded data #11b, and export header coded data #11a to header information lsb decoder 12 respectively, export MB coded data #11b to MB configuration part 13.

In header information lsb decoder 12, according to header coded data #11a, header information #12 is decoded.At this, header information #12 is the information of the size comprising input picture.

In MB configuration part 13, based on inputted header information #12, MB coded data #11b is separated into the coded data #13 corresponding with each macro block, and MB lsb decoder 14 is exported successively.

MB lsb decoder 14, by decoding successively to the inputted coded data #13 corresponding with each macro block, generates and exports the decoded picture #2 corresponding with each macro block.In addition, also decoded picture #2 is exported frame memory 15.The formation of MB lsb decoder 14 is by aftermentioned, and therefore, in this description will be omitted.

Carrying recorded decoding image #2 in frame memory 15.In frame memory 15, on the time point that specific macro block is decoded, record the decoded picture corresponding with the whole macro blocks be positioned under raster scan order before this macro block.

On the time point that the decoded picture generating process of the macro block unit that MB lsb decoder 14 performs for the whole macro blocks in image terminates, in moving image decoding apparatus 1, complete the generating process of the decoded picture #2 corresponding with inputted coded data.

(MB lsb decoder 14)

Below, the accompanying drawing replacing reference illustrates MB lsb decoder 14.

Fig. 2 is the block diagram of the formation representing MB lsb decoder 14.As shown in Figure 2, MB lsb decoder 14 possesses: sub-block cutting part 141, prediction residual lsb decoder 142, sub-block decoded picture generating unit 143, Prediction Parameters lsb decoder 144, predicted picture generating unit 145 and MB decoded picture generating unit 146.

The time point that sub-block cutting part 141 is transfused at the coded data #13 of macro block unit starts, and exports sub-block locations information #141a for representing each sub-block (each forecasting object region) position in this macro block being formed macro block (unit area) and the coded data relevant to the sub-block shown in sub-block locations information #141a and subblock coding data #141b successively with given order.In addition, the dividing method to the sub-block of macro block can be applied in the method generating and use in the dynamic image encoding device of coded data #1.

In addition, preferably sub-block cutting part 141 is set to following formation: after the sub-block locations information #141a relevant in the sub-block outputed to belong to 1st group described later and subblock coding data #141b, export the sub-block locations information #141a relevant to the sub-block belonging to 2nd group described later and subblock coding data #141b.Such as, be preferably set to following formation: scan the sub-block belonging to the 1st group with raster scan order, next, scan the sub-block belonging to the 2nd group with raster scan order.In addition, sub-block cutting part 141 can according to generating the identical order of the order that uses in the dynamic image encoding device of coded data #1, sub-block positional information #141a and subblock coding data #141b is exported.

Prediction residual lsb decoder 142, by the decoding of inputted subblock coding data #141b application variable length code, generates the conversion coefficient for sub-block of the sub-block locations information #141a that expression inputs.In addition, prediction residual lsb decoder 142, by being of a size of the inverse transformation of the DCT (discrete cosine transform) of same size to the sub-block shown in the conversion coefficient application generated and sub-block locations information #141a, generating and exports decoded residual #142.

Prediction Parameters lsb decoder 144, based on sub-block locations information #141a and subblock coding data #141b, is decoded and exports the Prediction Parameters #144 for each sub-block.

At this, Prediction Parameters refers to the parameter used in the generation of predicted picture.As the example of Prediction Parameters, include, for example the weight coefficient etc. in the predictive mode in infra-frame prediction, the motion vector in motion compensated prediction, luminance compensation prediction.

In addition, Prediction Parameters #144 comprises: the Prediction Parameters #43 exported from the 1st Prediction Parameters the lsb decoder 43 described later and Prediction Parameters #45 exported from the 2nd Prediction Parameters lsb decoder 45 described later.About the concrete formation of Prediction Parameters lsb decoder 144 and action by aftermentioned, therefore, in this description will be omitted.

Predicted picture generating unit 145, based on the decoded picture #15 recorded in Prediction Parameters #144, decoded picture #2 and frame memory 15, generates and exports the predicted picture #145 corresponding with forecasting object sub-block.About the concrete generation method of the predicted picture #145 in predicted picture generating unit 145 by aftermentioned, therefore, in this description will be omitted.

Sub-block decoded picture generating unit 143, by adding predicted picture #145 to decoded residual #142, generates and exports decoded picture and the sub-block decoded picture #143 of sub-block unit.

MB decoded picture generating unit 146 by accumulating the sub-block decoded picture #143 of sub-block unit by each macro block, and merges the whole sub-block decoded picture #143 forming macro block, generates and the decoded picture #2 of output macro block unit.The decoded picture #2 generated also is provided to predicted picture generating unit 145.

(Prediction Parameters lsb decoder 144)

Next, the formation of Prediction Parameters lsb decoder 144 is described with reference to Fig. 3.

Fig. 3 is the block diagram of the formation representing Prediction Parameters lsb decoder 144.As shown in Figure 3, Prediction Parameters lsb decoder 144 possesses: group detection unit 41, switching part 42, the 1st Prediction Parameters lsb decoder 43, reduced set leading-out portion 44 and the 2nd Prediction Parameters lsb decoder 45.

(group detection unit 41)

Which group in group detection unit 41 judge that the sub-block shown in sub-block locations information #141a belongs to prespecified multiple groups, and will represent that the group information #41 of result of determination exports switching part 42.

At this, above-mentioned prespecified multiple groups refer to, such as, classify to each sub-block in the dynamic image encoding device generating coded data #1 and multiple groups of obtaining.Namely, in the dynamic image encoding device generating coded data #1, belong to the sub-block SB1 ~ SBNs (Ns is the sum of the sub-block belonging to macro block MB) of certain macro block MB each be classified as to organize any one in GP1 ~ GPM (sub-block belonging to macro block MB is carried out the sum of sorted group by M) based on given sorting technique and sub-block SBn is classified as group GPm when, above-mentioned group of detection unit, such as based on above-mentioned given sorting technique, is judged to be that the sub-block SBn shown in sub-block locations information #141a belongs to group GPm.

Below, each sub-block to be categorized into the situation of 2 groups, be described with reference to Fig. 4 (a) ~ (f).

Fig. 4 (a) ~ (b) represents that 16 sub-blocks contained in macro block MB are classified as the figure of any one in the 1st group and the 2nd group separately based on sorting technique A, Fig. 4 (c) ~ (d) is the figure representing the situation that each sub-block is classified based on sorting technique B, Fig. 4 (e) ~ (f) is the figure representing the situation that each sub-block is classified based on sorting technique C.

Each sub-block contained in macro block MB, can as shown in Fig. 4 (a) ~ (b), the mode being checker flag shape according to the configuration of sub-block contained in each group is classified as the 1st group or the 2nd group, also can as shown in Fig. 4 (c) ~ (d), according to sub-block contained in each group only laterally adjacent mode be classified, can also as shown in Fig. 4 (e) ~ (f), according to sub-block contained in each group only the adjacent mode of longitudinal direction be classified.

Generally speaking, according to the relevant difference spatially of the Prediction Parameters in macro block, best sorting technique is different.Even if it is also effective sorting technique in case that any one the space correlation of above-mentioned sorting technique A in longitudinal direction and transverse direction is deposited.On the other hand, exist in this macro block oblique edge such, above-mentioned sorting technique B and sorting technique C is effective.

No matter use any sorting technique, all can be clear and definite from Fig. 4 (a) ~ (f), each sub-block is classified as any one in the 1st group and the 2nd group according to the position in macro block MB.

Group detection unit 41 reference sub-block positional information #141a, based on generating the sorting technique used in the dynamic image encoding device of coded data #1, is judging the sub-block shown in sub-block locations information #141a belongs to which group in the 1st group and the 2nd group.

Such as, in the dynamic image encoding device generating coded data #1, as shown in Fig. 4 (a) ~ (b), based on sorting technique A, the sub-block SB1 belonging to macro block MB be classified as the 2nd group and sub-block SB2 is classified as the 1st group when, group detection unit 41 based on this sorting technique A, by reference to sub-block locations information #141a, be judged to be that sub-block SB2 belongs to the 1st group, judge that sub-block SB1 belongs to the 2nd group.Also identical for other sub-block contained in macro block MB.

In addition, when using different sorting techniques by each macro block in the dynamic image encoding device generating coded data #1, preferably make coded data #1 comprise and represent that each macro block employs the mark of any sorting technique.By reference to such mark, even if when sorting technique is different by each macro block, group detection unit 41 also can carry out the judgement based on the sorting technique used in above-mentioned dynamic image encoding device.

In addition, although in the above description the number of sub-block contained in macro block has been set to 16, the present invention is not limited thereto (lower same).In addition, the sorting technique of the sub-block in macro block MB is not limited to above-mentioned example, can also be other sorting technique.Such as, can be the number of the sub-block belonging to the 1st group and the number such sorting technique different from each other (lower with) of sub-block belonging to the 2nd group.

(switching part 42)

The coded data relevant to the sub-block shown in sub-block locations information #141a and subblock coding data #141b, based on group information #41, are sent to any one parameter lsb decoder of the 1st Prediction Parameters lsb decoder 43 or the 2nd Prediction Parameters lsb decoder 45 by switching part 42.

Specifically, when being judged to be that the sub-block shown in sub-block locations information #141a belongs to the 1st group in group detection unit 41, switching part 42 sends above-mentioned subblock coding data #141b to the 1st Prediction Parameters lsb decoder 43, and when being judged to be that the sub-block shown in sub-block locations information #141a belongs to the 2nd group in group detection unit 41, switching part 42 sends above-mentioned subblock coding data #141b to the 2nd Prediction Parameters lsb decoder 45.

(the 1st Prediction Parameters lsb decoder 43)

1st Prediction Parameters lsb decoder 43, by carrying out the decoding of above-mentioned subblock coding data #141b, is decoded and exports the Prediction Parameters #43 used in the prediction of the sub-block (forecasting object sub-block) shown in sub-block locations information #141a in the dynamic image encoding device generating coded data #1.

More specifically, the 1st Prediction Parameters lsb decoder 43 first by use above forecasting object sub-block or in the prediction of the sub-block of left, the Prediction Parameters of having decoded, be set as the estimated value for this forecasting object sub-block.

Next, contained in the 1st Prediction Parameters lsb decoder 43 couples of subblock coding data #141b mark is decoded.

When this mark represent employ estimated value, above-mentioned estimated value is set as the Prediction Parameters for forecasting object sub-block, and when this mark represent do not use estimated value, the Prediction Parameters gone out from the partial decoding of h beyond above-mentioned mark is set as the Prediction Parameters for forecasting object sub-block.

In addition, when above forecasting object sub-block or the sub-block of left is not decoded, by above forecasting object sub-block or the decoding of left completes, the Prediction Parameters that uses in the prediction of the nearest sub-block of forecasting object sub-block, carry out reference as above-mentioned estimated value.

In addition, the Prediction Parameters #43 obtained that decodes also is provided to reduced set leading-out portion 44.

By above action, provide each sub-block decoding being subordinated to the 1st group and the Prediction Parameters #43 obtained to reduced set leading-out portion 44.

(reduced set leading-out portion 44)

Reduced set leading-out portion 44 accumulates Prediction Parameters #43, generates the Prediction Parameters collection RS (hereinafter referred to as " reduced set RS ") through reduction.At this, reduced set RS refers to, comprises the set of the Prediction Parameters #43 of each sub-block decoding being subordinated to the 1st group.In addition, the Prediction Parameters beyond Prediction Parameters #43 can also be comprised in reduced set RS.

In addition, when be subordinated in multiple sub-blocks of the 1st group decode identical Prediction Parameters such, reduced set leading-out portion 44, according to the mode this identical Prediction Parameters only being comprised to 1, generates reduced set RS.In other words, reduced set leading-out portion 44 generates reduced set RS according to the unduplicated mode of Prediction Parameters.Such as, when being subordinated to the sub-block SB1 ~ SB8 in the sub-block SB1 ~ SB16 of the 1st group and decoding Prediction Parameters PP1 respectively and decoding Prediction Parameters PP2 respectively from sub-block SB9 ~ SB16, reduced set leading-out portion 44 respectively generates reduced set RS containing the mode of 1 according to by Prediction Parameters PP1 and Prediction Parameters PP2.

Below, be the situation of the intra prediction mode H.264/MPEG-4AVC in specification for Prediction Parameters, the generation action of the reduced set RS performed by reduced set leading-out portion 44 is described with reference to Fig. 5 and Fig. 6 (a) ~ (c).

Fig. 5 represents the intra prediction mode (hereinafter referred to as " predictive mode ") and the figure to the call number that each predictive mode is given that use in the infra-frame prediction in H.264/MPEG-4AVC specification.It is the prediction direction used that each intra prediction mode is characterized in infra-frame prediction, as shown in Figure 5, in H.264/MPEG-4AVC specification, use the predictive mode in 8 directions (with call number 0,1,3 ~ 8 corresponding) and DC predictive mode (corresponding with call number 2).Below, the predictive mode of being specified by call number I is characterized by predictive mode I.In addition, the parameter set be made up of predictive mode 0 ~ predictive mode 8 is called basic parameter collection.

(the generation example 1 of reduced set RS)

Fig. 6 (a) is the flow chart of the 1st example of the generation action of the reduced set RS represented in reduced set leading-out portion 44.

As shown in Fig. 6 (a), first, reduced set leading-out portion 44, by reduced set RS is set as sky, carries out the initialization (step S101) of reduced set RS.

Next, the Prediction Parameters #43 that each sub-block being subordinated to the 1st group decodes is appended to (step S102) in reduced set RS by reduced set leading-out portion 44.Such as, when each sub-block being subordinated to the 1st group decode predictive mode 1, predictive mode 6 and predictive mode 8, predictive mode 1, predictive mode 6 and predictive mode 8 are appended in reduced set RS by reduced set leading-out portion 44.

By carrying out above action, in the 1st example, reduced set leading-out portion 44 can generate the reduced set RS that the Prediction Parameters #43 that decoded by each sub-block being subordinated to the 1st group is formed.

Generally speaking, exist relevant each other for the Prediction Parameters of the best for forming each sub-block of macro block.Therefore, be also that the possibility of best Prediction Parameters is high for belonging to each sub-block of the 1st group and the Prediction Parameters selected to each sub-block belonging to the 2nd group.In addition, the number of contained in above-mentioned reduced set RS predictive mode is less than the number that basic parameter concentrates contained Prediction Parameters.

Therefore, generate the dynamic image encoding device of coded data #1 by taking the formation corresponding with the formation of this example, can generated code amount is few under the prerequisite of not sacrificing code efficiency coded data #1.In addition, moving image decoding apparatus 1 is by taking the formation of this example, and coded data #1 that can be few to the code amount so generated decodes.

(the generation example 2 of reduced set RS)

Fig. 6 (b) is the flow chart of the 2nd example of the generation action of the reduced set RS represented in reduced set leading-out portion 44.

As shown in Fig. 6 (b), first, reduced set leading-out portion 44, by reduced set RS is set as sky, carries out the initialization (step S201) of reduced set RS.

Next, additional parameter set AS is appended to (step S202) in reduced set RS by reduced set leading-out portion 44.At this, in additional parameter set AS, preferably contain the Prediction Parameters being with the tendency be frequently used.Generally speaking, in the predictive mode of H.264/MPEG-4AVC specification, the predictive mode of being specified by less call number has more the tendency be frequently used in infra-frame prediction, therefore preferably containing the predictive mode of being specified by the call number that call number 0 ~ 8 is medium and small.Such as, preferably additional parameter set AS is configured to comprise predictive mode 0 (vertical direction predictive mode), predictive mode 1 (horizontal direction predictive mode) and predictive mode 2 (DC predictive mode).

In addition, can preferably additional parameter set AS be configured to comprise at least a kind of predictive mode in predictive mode 0, predictive mode 1 and predictive mode 2.

Next, the Prediction Parameters #43 that each sub-block being subordinated to the 1st group decodes is appended to (step S203) in reduced set RS by reduced set leading-out portion 44.Wherein, reduced set leading-out portion 44, for the Prediction Parameters be included in reduced set RS in Prediction Parameters #43, is not appended in reduced set RS for avoiding repetition.Such as, when Prediction Parameters #43 be predictive mode 1 and predictive mode 4 and in reduced set RS containing predictive mode 1 and predictive mode 2 such, reduced set leading-out portion 44 only adds predictive mode 4 in reduced set RS.

By carrying out above action, in the 2nd example, reduced set leading-out portion 44 can generate the reduced set RS that predictive mode contained in the Prediction Parameters #43 and additional parameter set decoded by each sub-block being subordinated to the 1st group is formed.

By constructing reduced set RS as described above, the reduced set RS that the Prediction Parameters #43 that decoded by each sub-block being subordinated to the 1st group and the predictive mode that there is the tendency be frequently used are formed can be generated.

Therefore, that generate coded data #1, that the code amount of the dynamic image encoding device energy generation forecast residual error of the reduced set leading-out portion that possesses action as this example is less coded data #1.In addition, the moving image decoding apparatus 1 possessing the reduced set leading-out portion 44 of action as this example can be decoded by the coded data #1 less to the code amount of such prediction residual.

(the generation example 3 of reduced set RS)

Fig. 6 (c) is the flow chart of the 3rd example of the generation action of the reduced set RS represented in reduced set leading-out portion 44.

As shown in Fig. 6 (c), first, reduced set leading-out portion 44, by reduced set RS is set as sky, carries out the initialization (step S301) of reduced set RS.

Next, the Prediction Parameters #43 that each sub-block being subordinated to the 1st group decodes is appended to (step S302) in reduced set RS by reduced set leading-out portion 44.

Next, reduced set leading-out portion 44 judges log ₂(Np-1) whether be integer (step S303).At this, Np is the number of Prediction Parameters contained in reduced set RS.

At log ₂(Np-1), when being integer ("Yes" of step S303), reduced set leading-out portion 44 exports reduced set RS.

At log ₂(Np-1), when not being integer ("No" of step S303), given Prediction Parameters is appended to (step S304) in reduced set by reduced set leading-out portion 44, and again carries out the process of step S303.At this, as above-mentioned given Prediction Parameters, such as, get basic parameter concentrate in contained predictive mode 0 ~ predictive mode 8, be not included in reduced set RS and the minimum predictive mode of call number.

As mentioned above, the predictive mode of being specified by less call number has by the tendency frequently used in infra-frame prediction.Therefore, reduced set leading-out portion 44 is appended to having in infra-frame prediction in reduced set by the predictive mode of the tendency frequently used in this step.

By carrying out above action, in the 3rd example, reduced set leading-out portion 44 can generate comprise Prediction Parameters #43 that each sub-block of being subordinated to the 1st group decodes, comprise 2 ⁿthe reduced set RS of+1 (n is integer) Prediction Parameters.

Generally speaking, when carrying out Variable Length Code together with mark whether identical with estimated value with expression for Prediction Parameters, the number of Prediction Parameters is 2 ⁿthe number that the situation of+1 (n is integer) compares to Prediction Parameters is not 2 ⁿ, there is the tendency that compression efficiency when carrying out Variable Length Code to Prediction Parameters is improved in the situation of+1.

Therefore, reduced set leading-out portion 44 by carrying out above-mentioned action, the reduced set RS that compression efficiency when can generate Variable Length Code is high.Therefore generate coded data #1, the moving image decoding apparatus of the reduced set leading-out portion that possesses action as this example can generate the high coded data #1 of compression efficiency.In addition, the moving image decoding apparatus 1 possessing the reduced set leading-out portion 44 of action as this example can be high to such compression efficiency coded data #1 decode.

In addition, not 2 in the number of Prediction Parameters #43 ⁿwhen+1 (n is integer), reduced set leading-out portion 44 according to comprising the mode of above-mentioned given Prediction Parameters to generate reduced set RS, therefore can generate the reduced set RS comprising and be with by the predictive mode of the tendency frequently used.

(the generation example 4 of reduced set RS)

In the generation example of kind of the reduced set RS of 3 shown in Fig. 6 (a) ~ (c), reduced set leading-out portion 44 is all that the Prediction Parameters of all categories after eliminating the Prediction Parameters of repetition in the Prediction Parameters #43 each sub-block being subordinated to the 1st group decoded is appended in reduced set RS.But can also be configured to: the Prediction Parameters decoded about the sub-block being subordinated to the 1st group, be not additional all categories but only add a part of kind.

Specifically, can be configured to: in the set of the Prediction Parameters decoded in the sub-block being subordinated to the 1st group, only will occur that the Prediction Parameters of ratio higher than set-point is appended in reduced set RS.At this, the appearance ratio of Prediction Parameters is such as by defining the number of the number being assigned the sub-block of this Prediction Parameters belonged in the sub-block of the 1st group divided by the whole sub-blocks belonging to the 1st group.Such as, if the number belonging to whole sub-blocks of the 1st group is Nf, and the number belonging in the sub-block of the 1st group, decode and be assigned Prediction Parameters Pa sub-block is Npa, then the appearance ratio of Prediction Parameters P defines by Npa/Nf.In addition, above-mentionedly occur that ratio also shows by percentage.

In addition, if the situation belonging to the 1st group with 8 sub-blocks (sub-block SB1 ~ SB8) is to further illustrate this generation example, then as follows.

Such as, when decoding predictive mode 0 to sub-block SB1, SB2, SB3, SB4, sub-block SB5, SB6 decoded to predictive mode 1, sub-block SB7 are decoded to predictive mode 2, when predictive mode 3 being decoded to sub-block SB8 and above-mentioned set-point is set as 40%, only by above-mentioned occur ratio be 50% predictive mode 0 be appended in reduced set RS.On the other hand, when above-mentioned set-point is set as 20%, by above-mentioned occur ratio be 50% predictive mode 0 and above-mentioned occur ratio be 25% predictive mode 1 be appended in reduced set RS.

Generally speaking, when the number of sub-block contained is in a macroblock many, the situation because being difficult to cut down efficiently code amount in reduced set RS containing multiple Prediction Parameters also can be produced.

According to this generation example, only by occurring that the Prediction Parameters of ratio higher than set-point is appended in reduced set RS in the set of the Prediction Parameters that reduced set leading-out portion 44 decodes in the sub-block being subordinated to the 1st group, the above-mentioned problem that can produce the situation being difficult to cut down code amount efficiently when the number of sub-block is many therefore can be solved.

Above, as illustrated in the generation example 1 ~ 4 of reduced set RS, reduced set RS can generate based on the Prediction Parameters belonging to the 1st group.More strictly say, reduced set RS at least can based on belong to the 1st group Prediction Parameters kind and belong to the 1st group each Prediction Parameters appearance ratio at least one and generate.

(the 2nd Prediction Parameters lsb decoder 45)

Next, the action of the 2nd Prediction Parameters lsb decoder 45 is described with reference to Fig. 7.In 2nd Prediction Parameters lsb decoder 45 pairs subblock coding data #141b contained among the coded data of each sub-block, in group detection unit 41, be judged to be that the Prediction Parameters P used in the prediction of each sub-block belonging to the 2nd group decodes.

In other words, 2nd Prediction Parameters lsb decoder 45 by information namely relevant with the Prediction Parameters for each sub-block belonging to the 2nd group for the information relevant to Prediction Parameters contained in subblock coding data #141b as a reference, to belong to the 2nd group each sub-block prediction in the Prediction Parameters P that uses decode.

In addition, the Prediction Parameters P decoded is exported as Prediction Parameters #45.

Fig. 7 is the flow chart of an example of the flow process of the decoding process represented in the 2nd Prediction Parameters lsb decoder 45.

As shown in Figure 7, first, the number N of contained in the 2nd Prediction Parameters lsb decoder 45 couples of reduced set RS Prediction Parameters counts (step S501).

Next, the 2nd Prediction Parameters lsb decoder 45 to judge in reduced set RS that the number N of contained Prediction Parameters is whether as 1 (step S502).

In a case of n=1 ("Yes" of step S502), unique Prediction Parameters contained in reduced set is set as Prediction Parameters P (step S503) by the 2nd Prediction Parameters lsb decoder 45.

When being not N=1 ("No" of step S502), the 2nd Prediction Parameters lsb decoder 45 derives Prediction Parameters estimated value Q (step S504).At this, Prediction Parameters estimated value Q refers to, the Prediction Parameters used in the prediction of the adjacent sub-block in the upside or left side with forecasting object sub-block.In addition, when the sub-block that the upside or left side with forecasting object sub-block is adjacent is not decoded, by above forecasting object sub-block or the decoding of left completes and the Prediction Parameters used in the prediction of the nearest sub-block of forecasting object sub-block is set to above-mentioned estimated value Q.

Next, the value decoded to representing that the mark whether Prediction Parameters of decoder object is identical with Prediction Parameters estimated value Q is decoded, and substitutes in variable a by the 2nd Prediction Parameters lsb decoder 45.

Below, be described for the situation that Prediction Parameters estimated value Q is identical with any one Prediction Parameters contained in reduced set RS.In addition, below, the value of variable a be 1 situation correspond to the Prediction Parameters of the decoder object situation identical with Prediction Parameters estimated value Q, the value of variable a be not the situation of 1 corresponding to the Prediction Parameters of the decoder object situation not identical with Prediction Parameters estimated value Q, be described as prerequisite.

Next, whether the value of the 2nd Prediction Parameters lsb decoder 45 decision variable a is 1 (step S506).

When the value of variable a is 1 ("Yes" of step S506), Prediction Parameters estimated value Q is set as Prediction Parameters P (step S507).

When the value of variable a is not 1 ("No" of step S506), whether the number N that the 2nd Prediction Parameters lsb decoder 45 to carry out in reduced set RS contained Prediction Parameters is the judgement (step S508) of 2.

When N=2 ("Yes" of step S508), the Prediction Parameters inconsistent with Prediction Parameters estimated value Q contained in reduced set RS is set as Prediction Parameters P (step S509) by the 2nd Prediction Parameters lsb decoder 45.

When being not N=2 ("No" of step S508), the 2nd Prediction Parameters lsb decoder 45 pairs ceil (log ₂(N-1)) Bit String of the length of bit is decoded, and the value decoded is substituted in variable b (step S510).At this, ceil (...) be by the value in parantheses more than integer in minimum integer be taken as the flow in upper plenum (lower with) of value.Therefore, ceil (...) value in parantheses is when being positive, can also show as and the value in parantheses is rounded up several functions.

Such as, in the case of n=5, the 2nd Prediction Parameters lsb decoder 45 couples of ceil (log ₂(5-1) Bit String of the length of)=2 bits is decoded, and the value decoded is substituted in variable b.At this, the value of variable b and the length of Bit String are that 2 bits are corresponding, get the arbitrary value in b=0,1,2,3.

Next, the Prediction Parameters in the Prediction Parameters inconsistent with Prediction Parameters estimated value Q contained in reduced set RS with (b+1) little call number is set as Prediction Parameters P (step S511) by the 2nd Prediction Parameters lsb decoder 45.

Such as, when the value of variable b is 0, the Prediction Parameters in the Prediction Parameters inconsistent with Prediction Parameters estimated value Q contained in reduced set RS with the 1st little call number is set as Prediction Parameters P by the 2nd Prediction Parameters lsb decoder 45.

In addition, the process illustrated in by step S504 and the Prediction Parameters estimated value Q that derives all different from arbitrary Prediction Parameters contained in reduced set RS, use the Prediction Parameters being endowed lowest index number in Prediction Parameters contained in reduced set RS to be used as above-mentioned Prediction Parameters estimated value Q.

It is more than an example of the decoding process performed by the 2nd Prediction Parameters lsb decoder 45.The Prediction Parameters P decoded by process as described above is exported as Prediction Parameters #45 by the 2nd Prediction Parameters lsb decoder.

As mentioned above, moving image decoding apparatus 1 be a kind of to by by the difference of original image and predicted picture, with represent have selected which Prediction Parameters in multiple Prediction Parameters of the generation method being used to specify predicted picture for each prediction unit selection information together with encode and moving image decoding apparatus that the coded data that obtains is decoded, it is characterized in that possessing: taxon (group detection unit 41), its by form multiple unit areas of predicted picture each in contained multiple prediction units be categorized as the 1st group or the 2nd group, 1st selected cell (the 1st Prediction Parameters lsb decoder 43), its with reference in above-mentioned selection information, for selection information i.e. the 1st selection information (information relevant to Prediction Parameters contained in subblock coding data #141b of each prediction unit belonging to the 1st group, the information relevant to the Prediction Parameters for each sub-block belonging to the 1st group), among the baseset be made up of prespecified Prediction Parameters, select the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of the 1st group, with the 2nd selected cell (the 2nd Prediction Parameters lsb decoder 45), it is with reference in above-mentioned selection information, for the selection information of each prediction unit belonging to the 2nd group, namely the 2nd information (the information relevant to Prediction Parameters contained in subblock coding data #141b is selected, the information relevant to the Prediction Parameters for each sub-block belonging to the 2nd group), from comprising the Prediction Parameters selected by above-mentioned 1st selected cell (the 1st Prediction Parameters lsb decoder 43) at least partially, and among the reduced set RS be made up of the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset, select the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of the 2nd group.

(other configuration example of Prediction Parameters lsb decoder 144)

Although describe reduced set leading-out portion 44 by each macro block to generate the formation of reduced set RS in the above-mentioned explanation relevant to Prediction Parameters lsb decoder 144, the present invention is not limited thereto.

In Prediction Parameters lsb decoder 144, such as can be configured to: reduced set leading-out portion 44 by each sub-block to generate reduced set RS, 2nd Prediction Parameters lsb decoder 45, based on the reduced set RS generated by each sub-block, is decoded for the Prediction Parameters of forecasting object sub-block.

In such formation, reduced set leading-out portion 44 as shown in Fig. 8 (a), by carrying out following process to generate reduced set RS.

(step S701)

First, reduced set leading-out portion 44, by reduced set RS is set as sky, carries out the initialization of reduced set RS.

(step S702)

Next, the region that the sub-block of the periphery by forecasting object sub-block is formed is set as neighbouring sub-block region NSR by reduced set leading-out portion 44.

Fig. 8 (b) is the figure of the example representing neighbouring sub-block region NSR.As shown in Fig. 8 (b), neighbouring sub-block region NSR such as can by forecasting object sub-block periphery, be that sub-block within 1 ~ 3 urban district distance is formed in units of sub-block time from the distance of forecasting object sub-block.At this, urban district distance is the distance that the absolute value sum of the difference of each coordinate by coordinate between 2 o'clock defines.

In addition, as shown in Fig. 8 (b), neighbouring sub-block region NSR generally can comprise the sub-block of the macro block belonged to beyond macro block belonging to forecasting object sub-block.

(step S703)

Next, the Prediction Parameters that the decoding among the Prediction Parameters for each sub-block contained in neighbouring sub-block region NSR completes is appended in reduced set RS by reduced set leading-out portion 44.

In addition, when contained in sub-block region NSR near identical Prediction Parameters is corresponding multiple sub-block, this identical parameter only adds 1 in reduced set RS by reduced set leading-out portion 44.

By carrying out above action, reduced set leading-out portion 44 can generate reduced set RS by each macro block.In addition, the 2nd Prediction Parameters lsb decoder 45 based on the reduced set RS generated by each sub-block, can be decoded for the Prediction Parameters of forecasting object sub-block.

Generally speaking, relevant for existing between the Prediction Parameters of forecasting object sub-block and the Prediction Parameters of the sub-block of the periphery for this forecasting object sub-block.Therefore, Prediction Parameters contained in the reduced set RS generated by above-mentioned process comprises the possibility of the optimal Prediction Parameters of prediction of the sub-block belonging to the 2nd group high.In addition, the number of Prediction Parameters contained in the reduced set RS generated by above-mentioned process is generally less than the number for the 1st group of optional Prediction Parameters.

Therefore, generate the dynamic image encoding device of coded data #1 by taking the formation corresponding with above-mentioned formation, can generated code amount is few under the prerequisite of not sacrificing code efficiency coded data #1.In addition, moving image decoding apparatus 1 is by taking above-mentioned formation, and coded data #1 that can be few to the code amount so generated decodes.

In addition, near multiple sub-blocks contained in the NSR of sub-block region, when there is not the Prediction Parameters of having decoded, the 2nd Prediction Parameters lsb decoder 45 is such as configured to select Prediction Parameters among basic parameter collection.

In addition, the reduced set 44 in this configuration example can be configured to: by the process almost same with the process illustrated in (the generation example 1 of reduced set RS) ~ (the generation example 4 of reduced set RS), derive reduced set RS.In the case, " the 1st group " in (the generation example 1 of reduced set RS) ~ (the generation example 4 of reduced set RS) is renamed as " the neighbouring sub-block region NSR " in this configuration example.

In addition, although be set to by reduced set RS in the above description the 2nd group of use, the present invention is not limited thereto.Above-mentioned process can be applied the whole sub-blocks in macro block.That is, can be configured to: for the whole sub-blocks in macro block, based on the reduced set RS generated by each sub-block, Prediction Parameters is decoded.

The dynamic image encoding device generating coded data #1, by taking the formation corresponding with above-mentioned formation, can cut down the code amount of the Prediction Parameters for the whole sub-blocks in macro block further.Therefore, the coded data #1 that above-mentioned dynamic image encoding device energy generated code amount is less.In addition, moving image decoding apparatus 1, by taking above-mentioned formation, can be decoded to the coded data #1 so generated.

(predicted picture generating unit 145)

Below, the generating process of the predicted picture #145 in predicted picture generating unit 145 is described.

The prediction direction of predicted picture generating unit 145 according to Prediction Parameters #144 (predictive mode), such as, generate the predicted pixel values of each pixel (forecasting object pixel) in predicted picture #145 (forecasting object sub-block) as follows.In addition, below, be that the situation of any one in the predictive mode 0 ~ predictive mode 8 shown in Fig. 5 is described for Prediction Parameters #144.

Predicted picture generating unit 145, after being assigned with the predictive mode shown in Prediction Parameters #144 to forecasting object pixel, carries out following action.

■ is when distributed predictive mode is beyond predictive mode 2 (DC prediction), the location of pixels of forecasting object pixel is set to starting point by predicted picture generating unit 145, and will that complete pixel towards the decoding on the reciprocal virtual line segment of prediction direction, be set to this forecasting object pixel from the pixel value of the nearest pixel (hereinafter referred to as nearest pixel) of this pixel pixel value be arranged in.In addition, the value that can calculate using the pixel value of the pixel of the pixel value of nearest pixel and the periphery of nearest pixel is set to the pixel value of this forecasting object pixel.

■ is predictive mode 2 and the sub-block adjacent with the upside of forecasting object sub-block is (following at distributed predictive mode, be called sub-block) and the sub-block adjacent with left side (following, be called left sub-block) be in when having decoded, the mean value of the pixel value of the pixel pixel value of the pixel of lower side a line of upper sub-block and the rightmost side one of left sub-block arranged is set to the pixel value of forecasting object pixel.

■ when distributed predictive mode be predictive mode 2 and on sub-block be in decoded and left sub-block be in decoding do not complete, by the pixel value of the pixel of lower side a line of upper sub-block and the left of forecasting object sub-block and the mean value of the pixel value of the pixel arranged from the rightmost side one in the nearest sub-block (hereinafter referred to as a left side closest to sub-block) of forecasting object sub-block is set to the pixel value of forecasting object pixel.

■ when distributed predictive mode be predictive mode 2 and on sub-block be in decoding do not complete and left sub-block be in decoded, by the top of forecasting object sub-block and from the nearest sub-block of forecasting object sub-block (hereinafter referred to as on closest to sub-block) in the pixel value of pixel of lower side a line and the mean value of the pixel value of pixel that arranges of the rightmost side one of left sub-block be set to the pixel value of forecasting object pixel.

■ when distributed predictive mode be predictive mode 2 and on sub-block and left sub-block be all in decoding do not complete, by to be set to the pixel value of forecasting object pixel closest to pixel value and the mean value of the pixel value of the left pixel arranged closest to the rightmost side one of sub-block of the pixel of lower side a line of sub-block.

Below, the example of the generating process of the predicted picture #145 performed by predicted picture generating unit 145 when forecasting object sub-block is 4 × 4 pixel is illustrated with reference to Fig. 9.

Fig. 9 is the figure of the pixel (reference pixel) of each pixel (forecasting object pixel) of the forecasting object sub-block representing 4 × 4 pixels and the periphery of this forecasting object sub-block.As shown in Figure 9, give label a ~ p to forecasting object pixel, give label A ~ M to reference pixel, and the pixel value of pixel X (X is a ~ p, in A ~ M any one) is characterized by X.In addition, complete if reference pixel A ~ M is all in decoding.

(predictive mode 0)

When distributed predictive mode is predictive mode 0, predicted picture generating unit 145 is by following formula pixel value in next life a ~ p

a，e，i，m＝A，

b，f，j，n＝B，

c，g，k，o＝C，

d，h，l，p＝D。

(predictive mode 2)

When distributed predictive mode is predictive mode 2 (DC prediction), predicted picture generating unit 145 is by following formula pixel value in next life a ~ p

a～p＝ave(A，B，C，D，I，J，K，L)。

At this, ave (...) represent element contained in parantheses is averaged.

(predictive mode 4)

When distributed predictive mode is predictive mode 4, predicted picture generating unit 145 is by following formula pixel value in next life a ~ p

d＝(B+(C×2)+D+2)＞＞2，

c，h＝(A+(B×2)+C+2)＞＞2，

b，g，l＝(M+(A×2)+B+2)＞＞2，

a，f，k，p＝(I+(M×2)+A+2)＞＞2，

e，j，o＝(J+(I×2)+M+2)＞＞2，

i，n＝(K+(J×2)+I+2)＞＞2，

m＝(L+(K×2)+J+2)＞＞2。

At this, " > > " represents dextroposition computing, value for arbitrary positive integer x, s, x > > s equals the value after removing the fractional part of x ÷ (2^s).

In addition, predicted picture generating unit 145, for the predictive mode beyond above-mentioned predictive mode, also carrys out calculating pixel value a ~ p by same method.

The remarks item > that < is relevant to moving image decoding apparatus

Although be illustrated for moving image decoding apparatus involved in the present invention above, the present invention is not limited to above formation.

(remarks item 1)

Such as, the 2nd Prediction Parameters lsb decoder 45 can be configured to: according to mark contained in coded data #1, switch in when decoding to Prediction Parameters #45 whether utilize reduced set RS.

More specifically, 2nd Prediction Parameters lsb decoder 45 such as can be configured to: the value of mark contained in coded data #1 is 1, reduced set RS is used to decode to Prediction Parameters #45, and the value of mark contained in coded data #1 is when being 0, replaces reduced set RS and use basic parameter collection to decode to Prediction Parameters #45.

By being set to such formation, the treating capacity when decoding to Prediction Parameters #45 can be reduced.

(remarks item 2)

In addition, the 2nd Prediction Parameters lsb decoder 45 can be configured to: according to the number of sub-block contained in macro block, switch in when decoding to Prediction Parameters #45 whether utilize reduced set RS.

More specifically, can be configured to: when the number of sub-block contained is in a macroblock more than 16, reduced set RS is used to decode to Prediction Parameters #45, and the number of sub-block contained is in a macroblock when being less than 16, replaces reduced set RS and use basic parameter collection to decode to Prediction Parameters #45.

By being set to such formation, treating capacity when Prediction Parameters #45 is decoded can be reduced.

(remarks item 3)

In addition, although the set employing Prediction Parameters as shown in Figure 5 is in the above description used as basic parameter collection, the present invention is not limited thereto.

Such as, Prediction Parameters lsb decoder 144 can be configured to: use the parameter set of attention vertical direction such shown in the parameter set of attention horizontal direction such shown in Figure 16 (a) or Figure 16 (b) to be used as basic parameter collection.

More specifically, Prediction Parameters lsb decoder 144 such as can be configured to: when macro block internal memory edge is in the horizontal direction such, the parameter set of attention horizontal direction such shown in Figure 16 (a) is used to be used as basic parameter collection, and when macro block internal memory edge is in the vertical direction such, use the parameter set of attention vertical direction such shown in Figure 16 (b) to be used as basic parameter collection.

In addition, can be configured to: when optionally using multiple like this basic parameter collection and the parameter set shown in Figure 16 (a) or (b) is chosen as basic parameter collection, 2nd Prediction Parameters lsb decoder 45 uses reduced set RS to decode to Prediction Parameters #45, and when the parameter set shown in Fig. 5 is chosen as basic parameter collection, the 2nd Prediction Parameters lsb decoder 45 replaces reduced set RS and uses this basic parameter collection to decode to Prediction Parameters #45.

By being set to such formation, reduced set RS can be utilized according to the characteristic of the image in macro block.

(dynamic image encoding device)

Below, with reference to Figure 10 ~ Figure 14, the dynamic image encoding device (picture coding device) 2 involved by present embodiment is described.Figure 10 is the block diagram of the formation representing dynamic image encoding device 2.As shown in Figure 10, dynamic image encoding device 2 possesses: header information determination section 21, header information coding unit 22, MB configuration part 23, MB coding unit 24, variable length code multiplexing unit 25, MB lsb decoder 26 and frame memory 27.

Dynamic image encoding device 2 in brief, is generate by encoding to input picture #100 and the device of outputting encoded data #1.

Header information determination section 21 decides header information based on input picture #100.The header information determined is output as header information #21.Picture size containing input picture #100 in header information #21.Header information #21 is not only input to MB configuration part 23, is also provided to header information coding unit 22.

Header information coding unit 22 couples of header information #21 encode, and complete header information #22 to coding and export.The header information #22 that encoded is provided to variable length code multiplexing unit 25.

Input picture #100, based on header information #21, is divided into multiple macro block by MB configuration part 23, and exports the macroblock image #23 relevant to each macro block.Macroblock image #23 is supplied to MB coding unit 24 successively.

MB coding unit 24 is encoded to by the macroblock image #23 inputted successively, generates MB coded data #24.The MB coded data #24 generated is provided to variable length code multiplexing unit 25.About the formation of MB coding unit 24 by aftermentioned, therefore, in this description will be omitted.

Variable length code multiplexing unit 25, by completing header information #22 to coding and MB coded data #24 carries out multiplexing, generates and outputting encoded data #1.

MB lsb decoder 26, by decoding successively to the inputted MB coded data #24 corresponding with each macro block, generates and exports the decoded picture #26 corresponding with each macro block.Decoded picture #26 is provided to frame memory 27.

The decoded picture #26 inputted is recorded in frame memory 27.On the time point that specific macro block is encoded, record the decoded picture corresponding with the whole macro blocks be positioned under raster scan order before this macro block.

(MB coding unit 24)

Below, the accompanying drawing replacing reference more specifically illustrates MB coding unit 24.

Figure 11 is the block diagram of the formation representing MB coding unit 24.As shown in figure 11, MB coding unit 24 possesses: sub-block cutting part 241, Prediction Parameters determination section 242, Prediction Parameters coding unit 243, prediction residual generating unit 244, transform coefficients encoding portion 245, prediction residual lsb decoder 246, sub-block decoded picture generating unit 247, predicted picture generating unit 248 and MB coded data generating unit 249.

Macroblock image #23 is divided into multiple sub-block by sub-block cutting part 241, and exports sub-block locations information #241a for representing the position of each sub-block in this macro block being formed macro block and the view data relevant to the sub-block shown in sub-block locations information #241a and sub-image #241b successively with given order.

In addition, preferably sub-block cutting part 241 is set to following formation: after the sub-block locations information #241a relevant in the sub-block outputed to belong to 1st group described later and sub-image #241b, export the sub-block locations information #241a relevant to the sub-block belonging to 2nd group described later and sub-image #241b.Such as, be preferably set to following formation: scan the sub-block belonging to the 1st group with raster scan order, next, scan the sub-block belonging to the 2nd group with raster scan order.

Prediction Parameters determination section 242 determines and exports the Prediction Parameters #242 used in the generation of the predicted picture relevant to the sub-block shown in sub-block locations information #241a.In addition, Prediction Parameters coding unit 243 couples of Prediction Parameters #242 encode, and output encoder Prediction Parameters #243.About the formation of Prediction Parameters determination section 242 and Prediction Parameters coding unit 243 by aftermentioned, therefore, in this description will be omitted.

Prediction residual generating unit 244 determines the sub-block as forecasting object based on sub-block locations information #241a, and generates in this sub-block, between sub-image #241b and the predicted picture #248 generated by predicted picture generating unit 248 difference image and prediction residual #244.

The prediction residual #244 application of 245 pairs, transform coefficients encoding portion and sub-block are of a size of the frequency translation of same size, generate the conversion coefficient of prediction residual #244.

In addition, transform coefficients encoding portion 245 is carrying out quantification to after generating quantization transform coefficient #245a to above-mentioned conversion coefficient, variable length code is generated to variable-length coding methods such as this quantization transform coefficient #245a application CABAC or CAVLC, and this variable length code is exported as coded data #245b.

Prediction residual lsb decoder 246 couples of quantization transform coefficient #245a carry out re-quantization, and thereafter, the inverse transformation (frequency inverse transformation) converted by applying frequency, is generated and export decoded residual #246.

In addition, the above-mentioned process performed by prediction residual generating unit 244, transform coefficients encoding portion 245 and prediction residual lsb decoder 246 does not limit the present invention.Such as, transform coefficients encoding portion 245 can omit said frequencies conversion and directly quantize prediction residual.

Sub-block decoded picture generating unit 247 can be crossed and be added with decoded residual #246 predicted picture #248, generates and exports sub-block decoded picture #247.

Predicted picture generating unit 248, based on Prediction Parameters #242, decoded picture #27 and sub-block decoded picture #247, generates and exports the predicted picture #248 corresponding with forecasting object sub-block.The concrete generation method of the predicted picture #248 in predicted picture generating unit 248 such as can apply the method same with the generation method of the predicted picture #145 in above-mentioned predicted picture generating unit 145.

MB coded data generating unit 249 by the accumulation coded data #245b relevant to each sub-block and the coded prediction parameter #243 relevant with each sub-block, and is merged into macro block unit, generates and the coded data of output macro block unit and MB coded data #24.

Below, replace the accompanying drawing of reference, Prediction Parameters determination section 242 and Prediction Parameters coding unit 243 are described.

(Prediction Parameters determination section 242)

Figure 12 is the block diagram of the formation representing Prediction Parameters determination section 242.As shown in figure 12, Prediction Parameters determination section 242 possesses: group detection unit 51, switching part 52, the 1st Prediction Parameters determination section 53, reduced set leading-out portion 54 and the 2nd Prediction Parameters determination section 55.

Sub-block shown in sub-block locations information #241a is categorized as arbitrary group in multiple groups by group detection unit 51, and switching part 52 is exported to the group information #51 of presentation class result.

Group detection unit 51, such as, like that, can be categorized as any one of the 1st group or the 2nd group by Fig. 4 (a) ~ (f) as already described by each sub-block.

In addition, group detection unit 51 can use different sorting techniques by each macro block, each sub-block is categorized as arbitrary group in multiple groups.Such as can be configured to: each sub-block forming macro block MB1 is categorized as 2 groups as shown in Fig. 4 (a) ~ (b), and each sub-block forming the macro block MB2 different from macro block MB1 is categorized as 2 groups as shown in Fig. 4 (c) ~ (d).So, when using any one sorting technique in multiple sorting technique to each macro block, group detection unit 51 is preferably made to export for representing the mark employing any sorting technique.By the moving image decoding apparatus of extremely being decoded to coded data #1 by this flag transmission, this moving image decoding apparatus can be identified in group detection unit 51 and employ any sorting technique.

The coded data relevant to the sub-block shown in sub-block locations information #241a and subblock coding data #241b, based on group information #51, are transferred to arbitrary parameter determination unit of the 1st Prediction Parameters determination section 53 or the 2nd Prediction Parameters determination section 55 by switching part 52.

Specifically, when the sub-block shown in sub-block locations information #241a is classified as the 1st group in group detection unit 51, switching part 52 sends above-mentioned subblock coding data #241b to the 1st Prediction Parameters determination section 53, and when being judged to the sub-block shown in sub-block locations information #241a to be categorized as the 2nd group in group detection unit 51, switching part 52 sends above-mentioned subblock coding data #241b to the 2nd Prediction Parameters determination section 45.

1st Prediction Parameters determination section 53, based on decoded picture #27, sub-block decoded picture #247 and subblock coding data #241b, decides (selection) and exports the Prediction Parameters #53 used in the generation of the predicted picture relevant to each sub-block belonging to the 1st group.In addition, Prediction Parameters #53 is also provided to reduced set leading-out portion 54.

Such as, when Prediction Parameters is the intra prediction mode H.264/MPEG-4AVC in specification, 1st Prediction Parameters determination section 53 for each sub-block belonging to the 1st group, select and export with the basic parameter collection shown in Fig. 5 to have illustrated among arbitrary predictive mode.

In addition, although the determining method of the concrete Prediction Parameters #53 in the 1st Prediction Parameters determination section 53 does not limit the present invention, the 1st Prediction Parameters determination section 53 such as decides Prediction Parameters #53 for each sub-block belonging to the 1st group according to the predicted picture in this sub-block and the minimum mode of the difference between input picture #100.Such as, 1st Prediction Parameters determination section 53 is set to and forms as follows: for the sub-block SB1 belonging to the 1st group, difference between the predicted picture generated at the predictive mode 1 used among basic parameter collection and input picture #100 minimum such, predictive mode 1 is selected to this sub-block SB1, and for belonging to the sub-block SB2 of the 1st group, difference between the predicted picture generated at the predictive mode 2 used among basic parameter collection and input picture #100 minimum such, predictive mode 2 is selected to this sub-block SB2.

Figure 13 (a) represents that the 1st Prediction Parameters determination section 53 is for belonging to each sub-block of the 1st group and the figure of the example of predictive mode that selects in each sub-block forming macro block MB.In the example shown in Figure 13 (a), to be have selected any one predictive mode in predictive mode 1, predictive mode 6 or predictive mode 8 for each sub-block belonging to the 1st group by the 1st Prediction Parameters determination section 53.In this example, predictive mode 1, predictive mode 6 and predictive mode 8 are provided to reduced set leading-out portion 54 as Prediction Parameters #53.

By above action, provide the Prediction Parameters #53 relevant to each sub-block belonging to the 1st group to reduced set leading-out portion 54.

The formation of reduced set leading-out portion 54 is identical with the reduced set leading-out portion 44 illustrated.That is, reduced set leading-out portion 54 usage forecastings parameter #53 generates reduced set RS.The generation method of the reduced set RS in reduced set leading-out portion 54 is identical with the generation method of the reduced set RS in the reduced set leading-out portion 44 illustrated.

In addition, when optionally using multiple generation method such in reduced set leading-out portion 54, reduced set leading-out portion 54 is preferably made to export for representing the mark that have selected any generation method.By the moving image decoding apparatus of extremely being decoded to coded data #1 by this flag transmission, this moving image decoding apparatus can be identified in reduced set leading-out portion 54 and employ any generation method.

Figure 13 (b) is the figure of the example representing the reduced set RS generated by reduced set leading-out portion 54 when having been carried out providing as Prediction Parameters #53 by each predictive mode shown in Figure 13 (a).As shown in Figure 13 (b), in this example, reduced set RS is made up of predictive mode 1, predictive mode 8 and predictive mode 6.

Select in the Prediction Parameters that 2nd Prediction Parameters determination section 55 is contained from reduced set RS and export the Prediction Parameters #55 used in the generation of the predicted picture relevant to each sub-block belonging to the 2nd group.

Although the determining method of the concrete Prediction Parameters #55 in the 2nd Prediction Parameters determination section 55 does not limit the present invention, the 2nd Prediction Parameters determination section 55 such as selects the Prediction Parameters #55 that the most suitably can generate the predicted picture in this sub-block in the Prediction Parameters contained from reduced set RS of each sub-block belonging to the 2nd group.

Figure 13 (c) is the figure of the example representing the predictive mode that each sub-block belonging to the 2nd group in each sub-block of the 2nd Prediction Parameters determination section 55 for formation macro block MB is selected from the reduced set RS shown in Figure 13 (b).As shown in Figure 13 (c), for each sub-block belonging to the 2nd piece, have selected any one predictive mode in predictive mode 1, predictive mode 6 or predictive mode 8 contained in the reduced set RS shown in Figure 13 (b).

Generally speaking, exist relevant each other for the Prediction Parameters of the best for forming each sub-block of macro block.Therefore, be also that the possibility of best Prediction Parameters is high for the Prediction Parameters selected by each sub-block belonging to the 1st group for each sub-block belonging to the 2nd group.

In addition, as mentioned above, select, for the Prediction Parameters #55 of each sub-block belonging to the 2nd group, therefore to compare to the situation not using reduced set RS in the Prediction Parameters that the 2nd Prediction Parameters determination section 55 is contained from reduced set RS, the code amount of Prediction Parameters #55 can be cut down.

Such as, as described later, when encoding together with mark whether identical with estimated value with expression for predictive mode, 3 kinds of predictive modes contained in the reduced set RS shown in Figure 13 (b) can use ceil (log ₂(3-1) the code amount of)=1 bit is encoded.On the other hand, when the 2nd Prediction Parameters determination section 55 does not use reduced set RS and select Prediction Parameters among the basic parameter collection be made up of 9 kinds of predictive modes, ceil (log will be needed ₂(9-1) the code amount of)=3 bits.In addition, in above-mentioned example, be made up of 8 sub-blocks due to the 2nd group, therefore by using reduced set RS, compare to the situation not using reduced set RS, the code amount of 3 × 8-1 × 8=16 bit can be cut down for each macro block.

Generally speaking, when encoding together with mark whether identical with estimated value with expression for Prediction Parameters, if the number of optional Prediction Parameters for each sub-block belonging to the 1st group is characterized by Nfs, the number of Prediction Parameters contained in reduced set RS is characterized by Nrs and the number of sub-block contained in the 2nd group is characterized by Ngs, then by using reduced set RS, compare to the situation not using reduced set RS, Ngs × (ceil (log can be cut down for each macro block ₂(Nfs-1))-ceil (log ₂(Nrs-1))) the code amount of bit.

So, by using reduced set RS, the code amount needed for coding of Prediction Parameters can be cut down under the prerequisite of not sacrificing code efficiency.

(other configuration example of Prediction Parameters determination section 242)

Although describe reduced set leading-out portion 54 by each macro block to generate the formation of reduced set RS in the above-mentioned explanation relevant to Prediction Parameters determination section 242, the present invention is not limited thereto.

Namely, following formation can be set in Prediction Parameters determination section 242: reduced set leading-out portion 54 by each sub-block to generate reduced set RS, 2nd Prediction Parameters determination section 55, based on the reduced set RS generated by each sub-block, decides the Prediction Parameters for forecasting object sub-block.

In such formation, reduced set leading-out portion 54 is configured to carry out the action same with the action of the reduced set leading-out portion 44 illustrated in (step S701) ~ (the step S703) of (other configuration example of Prediction Parameters lsb decoder 144).Wherein, the Prediction Parameters that the decoding in (step S701) ~ (the step S703) of (other configuration example of Prediction Parameters lsb decoder 144) completes, corresponding with the Prediction Parameters of having encoded in this example.

Thus, reduced set leading-out portion 54 can generate reduced set RS by each sub-block.In addition, the 2nd Prediction Parameters determination section 55 based on the reduced set RS generated by each sub-block, can decide the Prediction Parameters for forecasting object sub-block.

Therefore, dynamic image encoding device 1, can generated code amount is few under the prerequisite of not sacrificing code efficiency coded data #1 by taking above-mentioned formation.

In addition, when near multiple sub-blocks contained in the NSR of sub-block region there is not the Prediction Parameters of having encoded, the 2nd Prediction Parameters determination section 55 is such as configured to select Prediction Parameters among basic parameter collection.

In addition, the reduced set leading-out portion 54 in this configuration example can be configured to: by the process almost same with the process illustrated in (the generation example 1 of reduced set RS) ~ (the generation example 4 of reduced set RS), derive reduced set RS.In the case, " the 1st group " in (the generation example 1 of reduced set RS) ~ (the generation example 4 of reduced set RS) is renamed as " the neighbouring sub-block region NSR " in this configuration example.

In addition, although be set to by reduced set RS in the above description the 2nd group of use, the present invention is not limited thereto.Above-mentioned process can be applied the whole sub-blocks in macro block.That is, can be configured to: for the whole sub-blocks in macro block, based on the reduced set RS generated by each sub-block, decide Prediction Parameters.

By taking above-mentioned formation, the code amount of the Prediction Parameters for the whole sub-blocks in macro block can be cut down.Therefore, dynamic image encoding device 1, can the less coded data #1 of generated code amount by taking above-mentioned formation.

(Prediction Parameters coding unit 243)

Next, with reference to Figure 14, Prediction Parameters coding unit 243 is described.Figure 14 is the block diagram of the formation representing Prediction Parameters coding unit 243.As shown in figure 14, Prediction Parameters coding unit 243 possesses: group detection unit 61, switching part 62, the 1st Prediction Parameters coding unit 63, reduced set leading-out portion 64 and the 2nd Prediction Parameters coding unit 65.

Group detection unit 61 is formations almost same with the group detection unit 51 illustrated.That is, the sub-block shown in sub-block locations information #241a is categorized as arbitrary group in prespecified multiple groups by group detection unit 61, and is exported switching part 62 by the group information #61 of presentation class result.In addition, group detection unit 61 use and be identical sorting technique organizing the sorting technique used in detection unit 51, each sub-block is categorized as arbitrary group in prespecified multiple groups.

The Prediction Parameters #242 relevant to the sub-block shown in sub-block locations information #241a, based on group information #61, is sent to arbitrary parameter coding portion of the 1st Prediction Parameters coding unit 63 or the 2nd Prediction Parameters coding unit 65 by switching part 62.

Specifically, when the sub-block shown in sub-block locations information #241a is classified as the 1st group in group detection unit 61, above-mentioned Prediction Parameters #242 is sent to the 1st Prediction Parameters coding unit 63 and reduced set leading-out portion 64 by switching part 62, and when the sub-block shown in sub-block locations information #241a is classified as the 2nd group in group detection unit 61, above-mentioned Prediction Parameters #242 is sent to the 2nd Prediction Parameters coding unit 65 by switching part 62.

1st Prediction Parameters coding unit 63, by encoding to the Prediction Parameters #242 relevant to each sub-block belonging to the 1st group, generates and output encoder Prediction Parameters #63.

Specifically, the 1st Prediction Parameters coding unit 63 is first by for belonging to the sub-block of periphery of each sub-block of the 1st group and the Prediction Parameters selected is set as this sub-block estimated value.

Next, 1st Prediction Parameters coding unit 63 to representing that the mark whether Prediction Parameters selected for this sub-block is different from estimated value is encoded, and then is encoded to this Prediction Parameters when the Prediction Parameters selected for this sub-block is different from estimated value.

At this, when have selected the Prediction Parameters of each sub-block among basic parameter collection, Prediction Parameters comprises mark, and the code by 1 bit or 4 bits shows.

As mentioned above, by carrying out the coding employing estimated value, the compression ratio when encoding to the Prediction Parameters #242 relevant to each sub-block belonging to the 1st group can be improved.

In addition, the 1st Prediction Parameters coding unit 63 can be configured to the Prediction Parameters #242 relevant to each sub-block belonging to the 1st group directly to encode.

Reduced set leading-out portion 64 uses the Prediction Parameters #242 relevant to the sub-block belonging to the 1st group to generate reduced set RS.The generation method of the reduced set RS in reduced set leading-out portion 64 is identical with the generation method of the reduced set RS in the reduced set leading-out portion 44 illustrated, therefore, in this description will be omitted.

2nd Prediction Parameters coding unit 65 is by Prediction Parameters contained in reduced set RS, and namely for each sub-block belonging to the 2nd group, the Prediction Parameters selected is encoded, and generates and output encoder Prediction Parameters #65.

Specifically, the 2nd Prediction Parameters coding unit 65 is first by for belonging to the sub-block of periphery of each sub-block of the 2nd group and the Prediction Parameters selected is set as this sub-block estimated value.

Prediction Parameters number Nrs contained in reduced set RS is 1, does not carry out any coding and terminate the coded treatment of the 2nd Prediction Parameters.

Next, 2nd Prediction Parameters coding unit 65 to representing that the mark whether Prediction Parameters selected for this sub-block is different from estimated value is encoded, and then is encoded to this Prediction Parameters when the Prediction Parameters selected for this sub-block is different from estimated value.

At this, when Nrs is 2, terminate the coded treatment of the 2nd Prediction Parameters.

When in addition, Prediction Parameters contained in reduced set is by ceil (log ₂(Nrs-1)) code of bit shows.

In addition, the number Nrs of contained in reduced set RS Prediction Parameters is generally less than the number for the 1st group of optional Prediction Parameters.

Therefore, by using reduced set RS, can encode with less code amount for belonging to each sub-block of the 2nd group and the Prediction Parameters selected.

In addition, as mentioned above, by carrying out the coding employing estimated value, the compression ratio when encoding to the Prediction Parameters #242 relevant to each sub-block belonging to the 2nd group can be improved.

In addition, the 2nd Prediction Parameters coding unit 65 can be configured to: by for belonging to each sub-block of the 2nd group and the Prediction Parameters selected directly is encoded.

The remarks item > that < is relevant to dynamic image encoding device

Although be illustrated for dynamic image encoding device 2 involved in the present invention above, the present invention is not limited to above formation.

(remarks item 1 ')

Such as, the 2nd Prediction Parameters determination section 55 can be configured to: according to the size of the space correlation of Prediction Parameters in macro block, determines whether to utilize reduced set RS, when space correlation is little, does not utilize reduced set RS and determines Prediction Parameters #55.In addition, dynamic image encoding device 2 is preferably configured to: when the 2nd Prediction Parameters determination section 55 does not utilize reduced set RS and determines Prediction Parameters #55, to representing that the mark of unfavorable reduced set RS is encoded, and is sent to moving image decoding apparatus.

By being set to such formation, not using reduced set RS when the space correlation of Prediction Parameters is little and determine Prediction Parameters #55, therefore suppressing the treating capacity for determining Prediction Parameters.

(remarks item 2 ')

In addition, the 2nd Prediction Parameters determination section 55 can be configured to: according to the number of sub-block contained in macro block, switches in when determining Prediction Parameters #55 whether utilize reduced set RS.

More specifically, can be configured to: when the number of sub-block contained is in a macroblock more than 16, reduced set RS is used to decide Prediction Parameters #55, and the number of sub-block contained is in a macroblock when being less than 16, replaces reduced set RS and use basic parameter collection to decide Prediction Parameters #45.

(remarks item 3 ')

In addition, Prediction Parameters determination section 242 can be configured to: use the parameter set of attention vertical direction such shown in the parameter set of attention horizontal direction such shown in Figure 16 (a) or Figure 16 (b) to be used as basic parameter collection.

Such as, Prediction Parameters determination section 242 can be configured to: when macro block internal memory edge is in the horizontal direction such, the parameter set of attention horizontal direction such shown in Figure 16 (a) is used to be used as basic parameter collection, and when macro block internal memory edge is in the vertical direction such, use the parameter set of attention vertical direction such shown in Figure 16 (b) to be used as basic parameter collection.

Can be configured to: when optionally using multiple like this basic parameter collection and the parameter set shown in Figure 16 (a) or (b) is chosen as basic parameter collection, 2nd Prediction Parameters determination section 55 uses reduced set RS to decide Prediction Parameters #55, and when the parameter set shown in Fig. 5 is chosen as basic parameter collection, the 2nd Prediction Parameters determination section 55 replaces reduced set RS and uses this basic parameter collection to decode to Prediction Parameters #55.

By being set to such formation, reduced set RS can be utilized according to the characteristic of the image in macro block, therefore can cut down the code amount of Prediction Parameters efficiently.

(data structure of coded data #1)

Below, the data structure of the coded data #1 generated by dynamic image encoding device 2 is described with reference to Figure 15.

Figure 15 is the figure of the bit stream structure of the bit stream #MBS of each macro block of presentation code data #1.As shown in figure 15, bit stream #MBS comprises and contained sub-block SB1 ~ SBN (at this, N is the number of the sub-block in macro block) relevant information in macro block and sub-block information #SB1 ~ #SBN (at this, N is the number of the sub-block in macro block).

In addition, as shown in figure 15, each sub-block information #SBn (1≤n≤N) comprises: represent information and the sub-block locations information #Ln of the position of the sub-block SBn in macro block and represent the Prediction Parameters information #Pn establishing corresponding Prediction Parameters with sub-block SBn.

Sub-block locations information #Ln is the information in order to the position and reference of determining the sub-block SBn in macro block in the moving image decoding apparatus of decoding to coded data #1.Special in above-mentioned moving image decoding apparatus 1, sub-block locations information #Ln is the information in order to sub-block SBn being categorized as group and reference.

Prediction Parameters information #Pn is for determining to establish with sub-block SBn the information of corresponding Prediction Parameters in the moving image decoding apparatus of decoding to coded data #1.Special in above-mentioned moving image decoding apparatus 1, Prediction Parameters information #Pn represents the information for any one Prediction Parameters in the optional Prediction Parameters of the group belonging to sub-block SBn.

Such as, when sub-block SBn belongs to the 1st group and is basic parameter collection by the Prediction Parameters collection formed the 1st group of optional Prediction Parameters, Prediction Parameters information #Pn represents that basic parameter concentrates the information of any one predictive mode in contained predictive mode 0 ~ 8.

In addition, when the field boundary that sub-block SBn belongs to the 2nd group closes, Prediction Parameters information #Pn is the information of any one Prediction Parameters represented in reduced set RS in contained Prediction Parameters.

In addition, when coded data #1 is the data generated by encoding together with mark whether identical with estimated value with expression for predictive mode, Prediction Parameters information #Pn is by ceil (log ₂(N-1)) code of bit shows.At this, N is the number for the optional Prediction Parameters of the group belonging to sub-block SBn.

Generally speaking, the number of contained in reduced set RS Prediction Parameters is less than the number for the 1st group of optional Prediction Parameters.Therefore, the code amount of contained in coded data #1 Prediction Parameters is than few when not using reduced set RS.

Specifically, when coded data #1 is the data generated by encoding together with mark whether identical with estimated value with expression for predictive mode, the code amount of Prediction Parameters information #Pn compares to the situation not using reduced set RS, few ceil (log ₂(Nfs-1))-ceil (log ₂(Nrs-1)) bit.At this, Nfs is the number for the optional Prediction Parameters of each sub-block belonging to the 1st group, and Nrs is the number of Prediction Parameters contained in reduced set RS.

So, by using reduced set RS, the code amount of coded data #1 can be cut down.

In addition, sub-block locations information #Ln can not be contained in coded data #1.Such as, if when encoding and decoding time preset public sub-block scanning sequency, then can based on being that the such information of which sub-block in coded data #1 decides sub-block locations.

< is to the application examples > of other Prediction Parameters

Although main so that the predictive mode in infra-frame prediction is illustrated as Prediction Parameters in the above description, but the present invention is not limited thereto, generally speaking, also can be suitable for for other the parameter used during generation forecast image in the coded treatment/decoding process of moving image.

Below, to illustrate Prediction Parameters be the application examples of the situation of motion vector in motion compensated prediction and be the application examples of situation of weight coefficient in luminance compensation prediction to Prediction Parameters.

(application examples to motion vector)

In motion compensated prediction, use and be called that the Prediction Parameters of motion vector is to show that the decoding used in the prediction of forecasting object sub-block completes the position in the region on image.

Motion vector is selected among the Prediction Parameters collection depending on picture size and precision (interpolation precision) from the number of element.Such as, at the width (horizontal pixel count) of image for W, highly (longitudinal pixel count) for H and interpolation precision is 0.25 pixel, from the Prediction Parameters collection S defined by following formula, select motion vector V.

S≡(V|V＝((N/4)，(M/4))

At this, N, M are the integers of satisfied 0≤N < 4W, 0≤M < 4H.

When Prediction Parameters is such motion vector, the 1st Prediction Parameters determination section 53 such as decides and assigned motion vector by each sub-block belonging to the 1st group.In addition, the 1st parameter lsb decoder 43 can be decoded to motion vector by each sub-block belonging to the 1st group.

Reduced set leading-out portion 44,54 can from for belong to each sub-block of the 1st group and the motion vector distributed to generate reduced set RS.In addition, reduced set leading-out portion 44,54 can as describing, from the sub-block of the periphery for forecasting object sub-block and the motion vector distributed to generate reduced set RS.

For other each portion that moving image decoding apparatus 1 and dynamic image encoding device 2 possess, by carrying out the action same with when Prediction Parameters is predictive mode, the code amount of Prediction Parameters contained in coded data #1 can be cut down.

In addition, when Prediction Parameters is motion vector, reduced set leading-out portion 44,54 such as can be configured to: by make and for the sub-block of the periphery of forecasting object sub-block and the norm of the differential vector between the motion vector the distributed motion vector that to be below steady state value such be appended in reduced set RS.

Generally speaking, when Prediction Parameters is motion vector, Prediction Parameters collection comprises many predictive vectors.Such as, if establish W=2000, H=1000, then the number of contained in above-mentioned Prediction Parameters collection S motion vector is 8000 × 4000.

On the other hand, about the number of the motion vector accumulated in reduced set RS, the number belonging to the sub-block of the number of the sub-block of the 1st group or the periphery of forecasting object sub-block becomes the upper limit.Such as, even if employ sub-block contained in the neighbouring sub-block region NSR shown in Fig. 8 (b), the number of the motion vector accumulated in reduced set RS is at most 24.

Therefore, according to the characteristic of the image of coding/decoding object, there is following possibility: though only from the sub-block of the periphery for forecasting object sub-block and the motion vector distributed to generate reduced set RS, also the motion vector of sufficient amount cannot be accumulated in reduced set RS, therefore precision of prediction declines, and the code amount of residual error data increases.By making and for the sub-block of the periphery of forecasting object sub-block and the norm of the differential vector between the motion vector the distributed motion vector that to be below steady state value such is appended in reduced set RS, the code amount of Prediction Parameters can be cut down under the prerequisite not making the code amount of residual error data increase.

(application examples to the weight coefficient in luminance compensation prediction)

In luminance compensation prediction, each using the brightness of the multiple reference picture to the reference in order to motion compensated prediction of forecasting object sub-block is multiplied by weight coefficient and the value that obtains, predicts the brightness of this forecasting object sub-block.

The present invention can also apply situation Prediction Parameters being set to above-mentioned weight coefficient.Such as, reduced set leading-out portion 44,54 be configured to the sub-block from the periphery for forecasting object sub-block and the weight coefficient distributed to generate reduced set RS.

In addition, under situation of the present invention is applied to weight coefficient, can be configured to: each value of being got by weight coefficient is set up corresponding with the multiple typical values being prescribed sum in advance, and this typical value is used as Prediction Parameters.

Such as, when whole real number values (or value of set figure place) that the value W of weight coefficient can get satisfied 0≤W≤1 are such, can be configured to: the value Wn that will meet whole weight coefficients of (n/X)≤Wn≤((n+1)/X) sets up corresponding with typical value w, and typical value w is used as Prediction Parameters.At this, X is natural number, and n is the integer of satisfied 0≤n≤X-1.Corresponding by carrying out such foundation, the whole real number values meeting 0≤W≤1 can be mapped to the set of the typical value Wn ading up to X.In addition, the concrete value of X is such as set as that the code amount making coded data is less.

So, by the number of the element being used as Prediction Parameters is restricted to prespecified number, compares to the situation directly weight coefficient being used as Prediction Parameters, the code amount of coded data can be cut down.

(remarks item)

Picture coding device involved in the present invention is a kind of picture coding device of encoding to the difference of input picture and predicted picture, it is characterized in that comprising: taxon, predicted picture is divided into multiple unit area by it, and multiple prediction units contained in constituent parts region are categorized as the 1st group or the 2nd group; 1st selected cell, it is among the baseset be made up of prespecified Prediction Parameters, selects the Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 1st group; 2nd selected cell, it is among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset comprised the Prediction Parameters selected by above-mentioned 1st selected cell, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group; With Prediction Parameters coding unit, it have selected which Prediction Parameters and above-mentioned 2nd selected cell to above-mentioned 1st selected cell for each prediction unit belonging to above-mentioned 1st group and have selected which Prediction Parameters for each prediction unit belonging to above-mentioned 2nd group and encode.

According to above-mentioned such picture coding device formed, from comprising by above-mentioned 1st selected cell for belonging to the 2nd group among the reduced set formed at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset in each prediction unit of the 1st group comprised same unit area and the Prediction Parameters selected, select Prediction Parameters for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group, and which Prediction Parameters be have selected to above-mentioned 2nd selected cell encode.

In addition, preferably, above-mentioned reduced set only comprises that selected by above-mentioned 1st selected cell, different from each other whole Prediction Parameters.

According to above-mentioned formation, due to the number of Prediction Parameters contained in reduced set can be made less, therefore code amount can be cut down further.More specifically, due to being correlated with on the general Existential Space of Prediction Parameters, therefore create unbalanced in the distribution of the Prediction Parameters of the best for each prediction unit in specific region.So in specific region, the probability only utilized a part for various Prediction Parameters contained in baseset is high.Therefore, the set of the Prediction Parameters selected by the 1st selected cell in specific region compares to the set of Prediction Parameters contained in baseset, and the less situation of its yuan of prime number is more.Thus, by the set of the unduplicated whole Prediction Parameters selected by the 1st selected cell is set to reduced set, first prime number of reduced set can be reduced.

In addition, preferably, each of above-mentioned multiple unit area is the coding units in this picture coding device.

According to above-mentioned formation, owing to performing the generating process of above-mentioned reduced set once by each coding units (such as, the macro block H.264/MPEG-4AVC in specification), the treating capacity generated for reduced set therefore can be cut down.

In addition, preferably, in picture coding device involved in the present invention, the prediction unit belonging to above-mentioned 1st group and the prediction unit belonging to above-mentioned 2nd group are configured to grid checkerboard (checker flag shape).

Generally speaking, the Prediction Parameters in each prediction unit to exist between the Prediction Parameters in neighbouring prediction unit relevant.

According to above-mentioned formation, because the prediction unit belonging to above-mentioned 1st group is configured to grid checkerboard (checker flag shape) with the prediction unit belonging to above-mentioned 2nd group, be also therefore that the possibility of suitable Prediction Parameters is high for belonging to each prediction unit of above-mentioned 1st group and the Prediction Parameters selected to each prediction unit belonging to above-mentioned 2nd group.

Therefore, according to above-mentioned formation, the code amount needed for coding of Prediction Parameters can be cut down under the prerequisite of not sacrificing code efficiency.

In addition, above-mentioned Prediction Parameters may be used for specifying the predictive mode in infra-frame prediction.

According to above-mentioned formation, the code amount needed for coding of the predictive mode in infra-frame prediction can be cut down under the prerequisite of not sacrificing code efficiency.

In addition, preferably, above-mentioned 2nd selected cell, from comprising the Prediction Parameters selected by above-mentioned 1st selected cell and comprising among at least any one the reduced set in vertical direction predictive mode, horizontal direction predictive mode and the DC predictive mode infra-frame prediction, selects the Prediction Parameters for specifying the generation method of the predicted picture belonged in the prediction unit of above-mentioned 2nd group.

Generally speaking, vertical direction predictive mode, horizontal process predictive mode and DC predictive mode are predictive modes high by the frequency selected in infra-frame prediction.

According to above-mentioned formation, because above-mentioned 2nd selected cell is from comprising the Prediction Parameters selected by above-mentioned 1st selected cell and comprising among at least any one the reduced set in vertical direction predictive mode, horizontal direction predictive mode and the DC predictive mode infra-frame prediction, selecting the Prediction Parameters for specifying the generation method of the predicted picture belonged in the prediction unit of above-mentioned 2nd group, cutting down code amount with therefore making the reduction of the precision of prediction when generating the predicted picture belonged in the prediction unit of above-mentioned 2nd group.

In addition, can be configured to: the prediction units in the unit area comprising above-mentioned 1st group and above-mentioned 2nd group is more than prespecified threshold value, above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned reduced set, otherwise above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned baseset.

According to above-mentioned formation, due to Prediction Parameters can be selected among above-mentioned baseset the prediction units in unit area is less than set-point, the treating capacity for selecting Prediction Parameters therefore can be cut down.

In picture coding device involved in the present invention, can be configured to: above-mentioned basic energy collecting sets by each unit area, when meeting specific condition to the baseset set by the unit area comprising above-mentioned 1st group and above-mentioned 2nd group, above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned reduced set, otherwise above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned baseset.

According to above-mentioned formation, because above-mentioned basic energy collecting sets by each unit area, above-mentioned 2nd selected cell is when meeting specific condition to the baseset set by the unit area comprising above-mentioned 1st group and above-mentioned 2nd group, Prediction Parameters is selected among above-mentioned reduced set, otherwise among above-mentioned baseset, select Prediction Parameters, therefore while cutting down the treating capacity for selecting Prediction Parameters, code amount can be cut down.

In addition, picture coding device involved in the present invention can also show as: a kind of picture coding device of encoding to the difference of input picture and predicted picture, it is characterized in that, comprise: selected cell, among the reduced set at least partially of its Prediction Parameters of specifying from the generation method of the predicted picture the prediction unit comprised for completing the coding be arranged near this prediction unit, select the Prediction Parameters for specifying the generation method of the predicted picture in each prediction unit; With Prediction Parameters coding unit, which Prediction Parameters it have selected to above-mentioned selected cell for each prediction unit is encoded.

In addition, preferably, the code that the code length that above-mentioned Prediction Parameters coding unit uses ratio to represent that above-mentioned 1st selected cell have selected the code of which Prediction Parameters is short, is used as representing that above-mentioned 2nd selected cell have selected the code of which Prediction Parameters.

According to above-mentioned formation, because above-mentioned Prediction Parameters coding unit can use than representing the code that the above-mentioned 1st selected cell code length that have selected the code of which Prediction Parameters is short, be used as representing that above-mentioned 2nd selected cell have selected the code of which Prediction Parameters, therefore for each prediction unit belonging to above-mentioned 2nd group, the code that more short code is long can be used to encode to the information which Prediction Parameters expression have selected.

In addition, preferably, above-mentioned 2nd selected cell is from comprising the Prediction Parameters selected by above-mentioned 1st selected cell and by 2 ⁿamong the reduced set of+1 (n is arbitrary natural number) Prediction Parameters composition, select the Prediction Parameters for specifying the generation method of the predicted picture belonged in the prediction unit of above-mentioned 2nd group.

Generally speaking, to by 2 ⁿwhen the groups of elements of+1 (n is arbitrary natural number) element composition carries out encoding, compare to by 2 ⁿthe groups of elements of the element composition beyond+1 carries out the situation of encoding, and compression efficiency is improved.

According to above-mentioned formation, due to can from comprising the Prediction Parameters selected by above-mentioned 1st selected cell and by 2 ⁿamong the reduced set of+1 (n is arbitrary natural number) Prediction Parameters composition, selecting the Prediction Parameters for specifying the generation method of the predicted picture belonged in the prediction unit of above-mentioned 2nd group and encode to have selected which Prediction Parameters, therefore playing the further effect of the compression efficiency that can improve when encoding to Prediction Parameters.

According to above-mentioned such picture decoding apparatus formed, can from comprising by above-mentioned 1st selected cell for belonging to the 2nd group among reduced set form at least partially and by the Prediction Parameters below the number of Prediction Parameters contained in above-mentioned baseset in each prediction unit of the 1st group comprised same unit area and the Prediction Parameters selected, the Prediction Parameters of selection for specifying the generation method of the predicted picture belonged in each prediction unit of above-mentioned 2nd group.

According to above-mentioned formation, due to the number of Prediction Parameters contained in reduced set can be cut down further, therefore code amount can be cut down further.

In addition, preferably, each of above-mentioned multiple unit area is the decoding unit in this picture decoding apparatus.

According to above-mentioned formation, owing to performing the generating process of above-mentioned reduced set once by each decoding unit (such as, the macro block H.264/MPEG-4AVC in specification), the treating capacity generated for reduced set therefore can be cut down.

In addition, preferably, in picture decoding apparatus involved in the present invention, the prediction unit belonging to above-mentioned 1st group and the prediction unit belonging to above-mentioned 2nd group are configured to grid checkerboard (checker flag shape).

According to the picture coding device with the formation corresponding with above-mentioned formation, because the prediction unit belonging to above-mentioned 1st group is configured to grid checkerboard (checker flag shape) with the prediction unit belonging to above-mentioned 2nd group, therefore for each prediction unit belonging to above-mentioned 2nd group, suitable Prediction Parameters can be selected.Therefore, according to the picture coding device with the formation corresponding with above-mentioned formation, the code amount needed for coding of Prediction Parameters can be cut down under the prerequisite of not sacrificing code efficiency.

According to the picture decoding apparatus with above-mentioned formation, can the coded data after reducing code amount like this further be decoded.

In addition, in picture decoding apparatus involved in the present invention, above-mentioned Prediction Parameters may be used for specifying the predictive mode in infra-frame prediction.

According to above-mentioned formation, can decode to the coded data of the code amount of the predictive mode reduced in infra-frame prediction under the prerequisite of not sacrificing code efficiency.

According to the picture coding device with the formation corresponding with above-mentioned formation, because above-mentioned 2nd selected cell is from comprising the Prediction Parameters selected by above-mentioned 1st selected cell, and the vertical direction predictive mode comprised in infra-frame prediction, horizontal direction predictive mode, and among at least any one reduced set in DC predictive mode, select the Prediction Parameters for specifying the generation method of the predicted picture belonged in the prediction unit of above-mentioned 2nd group, therefore code amount is cut down with making the reduction of the precision of prediction when generating the predicted picture belonged in the prediction unit of above-mentioned 2nd group.

According to the picture decoding apparatus with above-mentioned formation, can the coded data few to such code amount decode.

In picture decoding apparatus involved in the present invention, can be configured to: above-mentioned basic energy collecting sets by each unit area, when meeting specific condition to the baseset set by the unit area comprising above-mentioned 1st group and above-mentioned 2nd group, above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned reduced set, otherwise above-mentioned 2nd selected cell selects Prediction Parameters among above-mentioned baseset.

According to above-mentioned formation, because above-mentioned basic energy collecting sets by each unit area, above-mentioned 2nd selected cell is when meeting specific condition to the baseset set by the unit area comprising above-mentioned 1st group and above-mentioned 2nd group, Prediction Parameters is selected among above-mentioned reduced set, otherwise among above-mentioned baseset, select Prediction Parameters, therefore while cutting down the treating capacity for selecting Prediction Parameters, the coded data after reducing code amount can be decoded.

In addition, picture decoding apparatus involved in the present invention can also show as: be a kind of difference to passing through input picture and predicted picture, with represent have selected which Prediction Parameters in multiple Prediction Parameters of the generation method being used to specify predicted picture for each prediction unit selection information together with to encode and the coded data that obtains carries out the picture decoding apparatus of decoding, it is characterized in that, comprise: selected cell, it is with reference to above-mentioned selection information, among the reduced set at least partially of the Prediction Parameters of specifying from the generation method of the predicted picture the prediction unit comprised for completing the decoding be arranged near this prediction unit, select the Prediction Parameters for specifying the generation method of the predicted picture in each prediction unit.

The picture decoding apparatus with above-mentioned formation can the coded data few to such code amount be decoded.

According to above-mentioned formation, can the coded data high to compression efficiency like this decode.

The present invention is not limited to above-mentioned each execution mode, can carry out various change in the scope that claim represents, carries out appropriately combined and execution mode that is that obtain is also contained in the scope of technology of the present invention to the technological means disclosed in different execution mode.

Industrial applicibility

The present invention preferably can be applied to and encode the picture decoding apparatus that the picture coding device that generates coded data and the coded data that generates using such picture coding device decode to image.

Label declaration

1 moving image decoding apparatus

14MB lsb decoder

144 Prediction Parameters lsb decoders

41 groups of detection units (taxon)

42 switching parts

43 the 1st Prediction Parameters lsb decoders (the 1st selected cell)

44 reduced set leading-out portions

45 the 2nd Prediction Parameters lsb decoders (the 2nd selected cell)

2 dynamic image encoding devices

24MB coding unit

242 Prediction Parameters determination sections

51 groups of detection units (taxon)

52 switching parts

53 the 1st Prediction Parameters determination sections (the 1st selected cell)

54 reduced set leading-out portions

55 the 2nd Prediction Parameters determination sections (the 2nd selected cell)

243 Prediction Parameters coding unit (Prediction Parameters coding unit)

248 predicted picture generating units

Claims

1. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess:

Prediction Parameters decoding unit, it is decoded to the Prediction Parameters being applied to each prediction unit;

Reduced set construction unit, it generates the reduced set as the set of Prediction Parameters; With

2nd Prediction Parameters decoding unit, it is to for determining among baseset that the 1st Prediction Parameters of contained Prediction Parameters is decoded, or, Bit String according to the length of the number decision by Prediction Parameters contained among above-mentioned reduced set is decoded to the 2nd Prediction Parameters, 2nd Prediction Parameters is for determining Prediction Parameters contained among reduced set

Above-mentioned reduced set comprise at least more than one for predicting the Prediction Parameters comprised in unit area near object prediction unit, and the number of Prediction Parameters contained among above-mentioned reduced set is less than the number of the Prediction Parameters contained by among the baseset that be made up of prespecified Prediction Parameters

Above-mentioned Prediction Parameters decoding unit is according to the value of pressing the mark contained by each prediction unit in coded data, the 1st Prediction Parameters gone out by above-mentioned 2nd Prediction Parameters decoding unit decodes is utilized to decode to above-mentioned Prediction Parameters, or utilize the 2nd Prediction Parameters gone out by above-mentioned 2nd Prediction Parameters decoding unit decodes to decode to above-mentioned Prediction Parameters, above-mentioned Prediction Parameters is the motion vector used in motion compensated prediction

When the number relevant to Prediction Parameters contained among above-mentioned reduced set is 1, the decoding that above-mentioned Prediction Parameters decoding unit omits the information relevant to Prediction Parameters is decoded to above-mentioned Prediction Parameters.

2. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess:

When the number relevant to Prediction Parameters contained among above-mentioned reduced set is 1, the decoding that above-mentioned Prediction Parameters decoding unit omits the information relevant to Prediction Parameters is decoded to above-mentioned Prediction Parameters,

Comprise among above-mentioned reduced set: for belong to above-mentioned near prediction unit area and belong to the Prediction Parameters of the prediction unit of the 2nd decoder object block different from the 1st decoder object block belonging to this prediction unit.

3. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess:

When for predict near object prediction unit there is the Prediction Parameters repeated between Prediction Parameters among at least plural Prediction Parameters that comprises in unit area, in the Prediction Parameters of above-mentioned repetition one is included in above-mentioned reduced set, above-mentioned Prediction Parameters decoding unit is decoded to Prediction Parameters contained among above-mentioned reduced set by each prediction unit

4. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess:

2nd Prediction Parameters decoding unit, it is to for determining among baseset that the 1st Prediction Parameters of contained Prediction Parameters is decoded, or, to for determining that the 2nd Prediction Parameters of Prediction Parameters contained among reduced set is decoded,

5. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess:

6. a picture decoding apparatus, to by by the difference of input picture and predicted picture, to encode together with the Prediction Parameters used in for the generation of each prediction unit at predicted picture and the coded data that obtains is decoded,

The feature of described picture decoding apparatus is to possess: